SAS And Reaction Wheels Are Useless

[Scarecrow71]

Is anybody else having issues keeping a trajectory?  In the last 2 hours, while trying to launch 1 rocket into orbit, I've experienced:

• Uncontrollable spinning during ascent
• Uncontrollable spinning after killing the engines
  • This includes while holding prograde and/or SAS hold
• Undesired/uncontrolled pitch/roll/yaw during maneuvers
  • This includes while holding prograde, maneuver, and/or SAS hold

This is becoming annoying.  A simple rocket, with 1 engine and 1 reaction wheel at the engine.  And it took me 2 hours to get that 1 rocket into orbit.  As far as I'm concerned, this needs to be the #1 priority with a patch; you simply cannot fly in this game without babysitting every last control.

 

/rant

[The Aziz]

Yep.

[Sitting Duck]

Yeah I found this out just now when using them for the first time. Disabling them in-flight has zero effect. Problem solved by removing them, which of course isn't supposed to be ideal.

If I let my Kerbals spin long enough, the capsule basically looked like it was being torn apart, so maybe it zero'd out all of the part GUIDs like in ShadowZone's video, here.
https://youtu.be/20kn1D6VKzQ?t=755

Another thing that will hopefully be fixed soon.

[Klapaucius]

Yes, I was trying to do the SSTO to the Mun challenge, and could not land my craft because I could not get it rolled over.  And when firing engines, it just went all over the place.  This should probably be logged in the bug section on the forum.

[schlosrat]

I'm not sure about useless, but I'd certainly go with "of dubious value while in atmo". Once you're out of atmo you'll certainly need them, and they seem to do about what you'd expect. I've been able to turn the pointing of even fairly large craft - one outside atmo and with a reasonable amount of reaction wheels. That said, the behavior of KSP2 in atmo is... umm... odd? Yeah... I'll go with odd.

What I've seen in KSP2 is that no matter what I do with TWR, reaction wheels, gimballing of motors, etc., in atmo things are just waaaaay more tippy and less stable. I've switched to using size MD Control Surfaces in places I'd ordinarily think I need at most a size SM Stabilizer. In fact, some of the "odd" behavior makes me think the problem is more about bugs in the atmo physics part than with the reaction wheels. Especially since reaction wheels seem to do OK and perform about like I'd expect out of atmo.

To make a good controllable rocket you want the CP to always be behind the CG. That's basic, right? So, if you take your design, drain all the fuel out of it, and find that the CG is *still* well ahead of CP then you might think, dang, this thing ought to be pretty stable! If part of that stability was some reasonably sized *Stabilizers*, then you might really expect the darn thing to do quite well on ascent. Nope.

Tear off those so-called stabilizers and replace them with Size Whatever *Control Surfaces*. These are basically like cannard wings - the whole darn things pivot, not just a flap. And if Size MD Stabilizers looked sane/rational/reasonable to you, then go with Size MD Control Surfaces. Oh, and when this bugger gets to Max Q watch them flip back and forth trying to control things! Your LV's will look more Buck Rodgers than NASA, but they may get out of the atmo a bit more reliably. I say *may* because, hey, this is KSP2 EA, and *may* is about all I'd promise even with this.

The only other tip I'd offer is make your gravity turn a bit later. On a well behaved LV I usually start my turn around 100 m/s and have hit 45 degrees by the time I'm at 10K m AGL. On a troublesome LV that really wants to flip I delay the gravity turn quite a bit and don't even get to 45 degrees until I'm well into the upper atmosphere. It's just nuts. Things should not be like that - and it's frankly antithetical to the whole "KSP2 will be easier for new players" thing. Nope! Not yet, anyway...

[LameLefty]

In the rockets I've been designing, SAS has been fine WHEN IT WORKS. During several hours this afternoon, at one point reaction wheels and SAS just stopped working at all. The only cure for that was reloading an earlier quicksave, after which they worked again. And other time, they were working right up until I had the rocket point towards a planned maneuver node, then the spinning started, an absolutely uncontrollable and increasing roll rate. Reloading quicksaves didn't help. I had to quit the game entirely and reload. After that, the same quickload files I had tried earlier were unaffected. 

Edited March 4, 2023 by LameLefty

[Chilkoot]

I run into this problem if the craft is under acceleration during a save.  When I load that save, the reaction wheels seem to become useless (sporadic).

I've also seen this problem surface after firing up the Labradoodle engine on an upper stage.   Sometimes* I get a nice, stable burn, other times it starts spinning uncontrollably, even at 10% thrust and a ton of reaction wheel built into the craft.

By "sometimes", I mean the same craft, so it doesn't appear to be related to the build.  Same rocket, sometimes it works, sometimes it doesn't.

[Scarecrow71]

Ok, what in the H-E-Double-Hockey-Sticks am I doing wrong here?  I see all these threads with people landing on other planets...and I can't even get into orbit due to SAS controls.  What the frak?  I'm frustrated and ready to shelve this whole thing.

[Scarecrow71]

I still want someone to acknowledge that SAS is broken and doesn't hold.  Either that or let me know what I'm doing wrong that everyone else is able to play.

[Bej Kerman]

4 minutes ago, Scarecrow71 said:

I still want someone to acknowledge that SAS is broken and doesn't hold.  Either that or let me know what I'm doing wrong that everyone else is able to play.

Perhaps we don't know enough about your current situation to determine a cause or workaround to your issue.

[LameLefty]

17 minutes ago, Scarecrow71 said:

I still want someone to acknowledge that SAS is broken and doesn't hold.  Either that or let me know what I'm doing wrong that everyone else is able to play.

SAS has worked fine for me, most of the time. For my use-case (rockets and spacecraft, not aircraft or rovers), I’ve found it tends to work just fine during ascent and most of the time in space. 

I have had one instance in ~30+ hours of play when SAS just simply stopped working entirely. I had to save, quit the game and reload and then it worked  again. 

The other SAS failing for me is failing to follow/track the maneuver node. That happened to me most recently this past Saturday when I had a very light probe and a frankly over-powered core stage trying to boost it to interplanetary speeds. I  resolved that by limiting the throttle of the engine to 50% and the gimbal range to 30%. After that, SAS held the pointing accurately though that burn and subsequent course corrections. 

Edited March 6, 2023 by LameLefty

[Scarecrow71]

16 minutes ago, LameLefty said:

SAS has worked fine for me, most of the time. For my use-case (rockets and spacecraft, not aircraft or rovers), I’ve found it tends to work just fine during ascent and most of the time in space. 

I have had one instance in ~30+ hours of play when SAS just simply stopped working entirely. I had to save, quit the game and reload and then it worked  again. 

The other SAS failing for me is failing to follow/track the maneuver node. That happened to me most recently this past Saturday when I had a very light probe and a frankly over-powered core stage trying to boost it to interplanetary speeds. I  resolved that by limiting the throttle of the engine to 50% and the gimbal range to 30%. After that, SAS held the pointing accurately though that burn and subsequent course corrections. 

I can't maintain a gravity turn without flipping, and I can't do burns to get into orbit because SAS will not hold.  I spin uncontrollably all over the place.  Simple probes.  Do I just have too much rocket?

[LameLefty]

Just now, Scarecrow71 said:

I can't maintain a gravity turn without flipping, and I can't do burns to get into orbit because SAS will not hold.  I spin uncontrollably all over the place.  Simple probes.  Do I just have too much rocket?

This should probably go in the Gameplay subforum, but yes.

The real world has nothing like “SAS” to hold rockets stable during launch. and it appears KSP is going to be more like the real world for launchers than KSP1’s over-powered SAS. 

Design your rockets to be less “long-and-heavy-at-the-end”, or else start your gravity turn way later, outside most of the atmosphere. Long and heavy-nosed rockets will (and ways have in the real world) required fins for pitch control.  Engine gimbaling can only do so much.  Too bad KSP2 doesn’t yet have something like KSP1’s aero-forces visualization. If it did and you put fins on your rocket, you’d see how much force was being required to counteract the pitching moment from all the mass at the payload end. 

[Scarecrow71]

1 minute ago, LameLefty said:

This should probably go in the Gameplay subforum, but yes.

The real world has nothing like “SAS” to hold rockets stable during launch. and it appears KSP is going to be more like the real world for launchers than KSP1’s over-powered SAS. 

Design your rockets to be less “long-and-heavy-at-the-end”, or else start your gravity turn way later, outside most of the atmosphere. Long and heavy-nosed rockets will (and ways have in the real world) required fins for pitch control.  Engine gimbaling can only do so much.  Too bad KSP2 doesn’t yet have something like KSP1’s aero-forces visualization. If it did and you put fins on your rocket, you’d see how much force was being required to counteract the pitching moment from all the mass at the payload end. 

Ok, here's a video I did this morning.  Please look at this and provide some feedback.  The game is unplayable for me right now.

 

[Bej Kerman]

10 minutes ago, Scarecrow71 said:

Ok, here's a video I did this morning.  Please look at this and provide some feedback.  The game is unplayable for me right now.

 

That looks like a lack of torque. Put RCS thrusters or lots of reaction wheels on.

[Scarecrow71]

6 minutes ago, Bej Kerman said:

That looks like a lack of torque. Put RCS thrusters or lots of reaction wheels on.

See, when I use reaction wheels it gets worse.  But I will try RCS.  Never had to do that before, though.

[Starhawk]

This discussion regarding problems with SAS control has been split off into it's own thread.

 

[thewhitemetroid]

I've had this problem with the labradoodle in second stage as well, along with other SAS abnormalities. I haven't been playing as much as you guys so I haven't experienced as many frustrating moments but I have seen enough to know the game is close to broken...not quite, but close. It doesn't make much sense right now to wonder what we are doing wrong. The game isn't ready. Just need some patience and wait for first patch to (hopefully) help smooth out these game breaking problems. Fingers crossed.

[modus]

1 hour ago, Scarecrow71 said:

Ok, here's a video I did this morning.  Please look at this and provide some feedback.  The game is unplayable for me right now.

 

I think the last time you would have made orbit if you kept sas on hold maneuver (or whatever it's called).

That said, I've experienced everything you showed in that video. 

I do keep my throttle a littler lower and my gravity turn less aggressive.

 

[Scarecrow71]

Well, spamming RCS thrusters helps with maneuver nodes.  It's just frustrating that I have zero problems in KSP1, but in KSP2 I can barely get to orbit.  And they think this is going to help new players want to play the game when veterans have these issues?

[Kerbart]

Maybe work your way up from less ambitious vessels? I get to orbit all the time, mind you the orbit  had in mind. I think it's easier to expand and see where things stop working (ad perhaps find out why) than launching large craft tah remain out of control no matter what you put on them.

Also: I noticed that sometimes directional controls don't work at all, but translation controls (IJKL) works instead.

[Scarecrow71]

38 minutes ago, Kerbart said:

Maybe work your way up from less ambitious vessels? I get to orbit all the time, mind you the orbit  had in mind. I think it's easier to expand and see where things stop working (ad perhaps find out why) than launching large craft tah remain out of control no matter what you put on them.

Also: I noticed that sometimes directional controls don't work at all, but translation controls (IJKL) works instead.

My problem here is that what I am used to in KSP1 doesn't work in KSP2, and it's frustrating.  I can build the same vessel in both versions of the game, and while I can get the thing to orbit in KSP1 with no issues, getting it there in KSP2 is hard.  Like, fighting the controls hard.  And I'm not talking about overly-complex vehicles, either. Simple, 2 or 3 stage apollo-style rockets.  And again, making the controls harder is not the way to get new players into the game.

[Starhawk]

And now that I check back on it, I see that the posts from Grand Discussion that I split off should have been merged into this thread.
That has now been done.

As regards the issues highlighted in your video I'll add my two cents worth.

First, some basic stuff.  Aside from getting out and pushing (i.e. EVAing your kerbal and moving the ship around with your jetpack) there are only three ways to turn your ship once it is in vacuum:

1. Reaction wheels
2. RCS thrusters
3. Engine gimbal (i.e. firing the engine at very low thrust and using the gimbal to turn the ship).

Reaction wheels seem to behave slightly differently in KSP2 compared to KSP1.  This is purely subjective and not tested in any way, but it seems that the ability of small reaction wheels to control very large masses drops off much more quickly than in KSP1.  In KSP1 reaction wheels where seriously overpowered and this may be a matter of giving them slightly more realistic values.  In any case, even in KSP1 the ship shown in the the youtube video I would absolutely expect to be uncontrollable with the tiny reaction wheel in the probe core.

More or larger reaction wheels can work very well to control a vehicle in vacuum.  The danger is if the ship is quite bendy the engine may be pointing in a slight off-line compared to the probe core.  This disparity in the direction the ship is actually being pushed and the direction that the probe core points can end up in an amplifying oscillation where the ship wobbles back and forth more and more violently and then explodes.  This behaviour could be seen in KSP1 as well.  The fix for this is to use fins or engine gimbal while in atmosphere and leave the reaction wheels turned off until you enter vacuum.

RCS tends to be more forgiving since it inherently has less thrust when the atmosphere is thickest and there is less fighting between engine gimbal direction and probe core direction.

Happy landings!

[schlosrat]

@Scarecrow71 Nice video, that does a great job of showing us what you’re dealing with. I think you’re very close with this design. I’d add one reaction wheel at the same size as your center stage and then tack on some control surfaces (not stabilizers) The control surfaces can help a lot while in atmo, and the reaction wheels are essential when out of it unless you want to putt around with RCS. I mainly only use rcs for docking, but as indicated by others, it can certainly be used to point while out of atmo.

 

Edited March 7, 2023 by schlosrat

[Browny1991]

I am literally having the same  issue as you @Scarecrow71 Ive used reaction wheels and RCS thrusters and the rocket wont keep a maneuver node . Ive put 8 rcs thusters and none of them are activiating to stabalise the rocket  i even tried manually activating them in parts manager but  nothing works .