Have to disable roll control on launching rockets in 1.0.5?

[Tfin]

I know how to fix it, since I can turn it off. The question is, is this a feature, or a bug, or what?

To explain:

I build a rocket, and it needs fins to gain enough control, as the capsule's reaction wheels are insufficient and I haven't unlocked other control methods yet in this save. (Not using the Swivel, of course. Workarounds not needed.)

Shortly after launch, it begins spinning. Not tumbling, it stays on course, but spinning. Rolling. The pilot is applying roll, I can see the angled fins. Turning off roll control for the parts fixes it, and doing so before launch prevents it from starting, so there isn't some other part causing it.

This didn't happen in earlier versions, as I never had to mess with this feature back then.

Is the pilot misinterpreting the controls and using them backwards now, or is it something else?

[Temstar]

What type of fin are you using?

[Tfin]

One with control surfaces, of course. Actually, this is with the... (go get the name...) AV-R8, but it doesn't always happen, which is why it catches me off-guard, and just clicking the "don't roll" button will stop it. I'd know it was this part if it were consistent or if it happened with roll control turned off.

[Freshmeat]

I think your pilot tries to stabilise the rocket. If you use radial decouplers, the boosters will tend to twist a bit, causing roll. The solution to that problem is to apply two struts from the booster to the rocket, going horizontally and in a v-shape. Try it and see if it helps.

[Frybert]

Well in a stock install the pilot should be you. So I'm wondering, are you using an autopilot? MechJeb perhaps?

[Archgeek]

1 hour ago, Frybert said:

Well in a stock install the pilot should be you. So I'm wondering, are you using an autopilot? MechJeb perhaps?

'Pretty sure Tfin means the SAS input from their kerbal pilot, there.

[Frybert]

59 minutes ago, Archgeek said:

'Pretty sure Tfin means the SAS input from their kerbal pilot, there.

Wow I completely forgot that SAS is tied to the pilot now. Well then, carry on.

[Kenchikuka]

The command is upside down?

I have problems like this with a probe upside down, after leave the atmosphere it stopped, in this case i dont have any kerbal on board.

Edited January 6, 2016 by Kenchikuka

[tomf]

If you are using radial boosters with control surfaces attached then is can cause uncontrolled spinning. The rocket starts spinning and the SAS tries to correct, but all that does is twist the boosters a little causing them to thrust asymetically and the spinng to accelerate. As mentioned above stuts to stop the boosters rotating can help. 

[sal_vager]

Maybe if we saw the craft we could help more, you can put a pic of it on imgur.com

[fourfa]

This sounds a lot like what happens when you build a rocket in the default north-facing orientation in the VAB, then launch with mechjeb.  Default options will roll to face east.

[Bandus]

I agree with folks who are speculating it is related to radial decouplers and no struts. I had rockets doing precisely the same thing before adding the struts.

[Frybert]

5 minutes ago, Bandus said:

I agree with folks who are speculating it is related to radial decouplers and no struts. I had rockets doing precisely the same thing before adding the struts.

OP seems to be indicating that the fins are causing the roll. He can see them deflecting and when he disables them roll stops.

 

OP look below your staging and you should see input on the 3 axis. Are any of them off center? If so, SOMETHING (sas, an autopilot or even a stuck key) is telling the ship to roll.

[Bandus]

2 minutes ago, Frybert said:

OP seems to be indicating that the fins are causing the roll. He can see them deflecting and when he disables them roll stops.

 

OP look below your staging and you should see input on the 3 axis. Are any of them off center? If so, SOMETHING (sas, an autopilot or even a stuck key) is telling the ship to roll.

 

Indeed. That would jive if SAS were active. The fins would be trying to compensate for the roll, which was being caused by no struts. Turning off the fin control surfaces would render them invalid and SAS, while still active, would no longer try to use them.

Edited January 6, 2016 by Bandus

[Frybert]

14 minutes ago, Bandus said:

 

Indeed. That would jive if SAS were active. The fins would be trying to compensate for the roll, which was being caused by no struts. Turning off the fin control surfaces would render them invalid and SAS, while still active, would no longer try to use them.

But if SAS was correcting for a booster induced roll, and the fin is disabled, shouldn't it start trying to roll the other way (the direction of the roll SAS was trying to counter)?

[Bandus]

57 minutes ago, Frybert said:

But if SAS was correcting for a booster induced roll, and the fin is disabled, shouldn't it start trying to roll the other way (the direction of the roll SAS was trying to counter)?

It could be so, unless the fins weren't enough to stop the booster caused spin in the first place. In that instance, it'd still roll the same way. However, I think the OP said that when the fin was disabled it stopped rolling. Admittedly, it seems a bit ambiguous to me. Without a screenshot or a .CRAFT file, it's hard to say what is happening here.

[Plusck]

Have you accidentally applied roll trim - using Alt-Q or Alt-E ?

 

[Snark]

+1 to please post a screenshot.  Whole lotta speculation going on in here that would likely be settled immediately if we could see what kind of ship you have. 

[Archgeek]

'Sounds like indeed a booster-induced roll worsened by SAS input on the fins.  With the fin surfaces disabled, SAS may well be managing to counter the roll with the command pod's reaction wheel.

Whether it be trim or SAS input, there should be a visible roll input in the control input UI.  If it's something weirder then you'll see the roll needle centered.

[JohnnRico]

I've got the very same problem here! In my case it is independent from my rocket. 

While in Space, SAS is turned off. Those little indicators on the lower left side wiggle a bit without me pushing any buttons. If I turn SAS on, everything is fine. If I choose the Maneuver Node at the Direction Selector, the ship will roughly point in the right direction, but will spin a bit and will not directly point at the node.

First, I thought it was a problem with my trimming. So, I turn off SAS and ALT-X ... which cause my roll indicator to switch to max roll. (And my ship will start to roll uncontrollably) Turning on SAS again will stabilize it. I could trim it then manually using ALT+q (or +e), but Alt-X will set the roll again to the max.

Any advice?

 

[Snark]

4 hours ago, JohnnRico said:

Any advice?

Yes.  "Post a screenshot, please." 

[Workable Goblin]

On January 7, 2016 at 10:06 AM, Archgeek said:

'Sounds like indeed a booster-induced roll worsened by SAS input on the fins.  With the fin surfaces disabled, SAS may well be managing to counter the roll with the command pod's reaction wheel.

Whether it be trim or SAS input, there should be a visible roll input in the control input UI.  If it's something weirder then you'll see the roll needle centered.

I disagree, because I've seen the same thing in rockets without any boosters at all. I built a simple single-stage rocket using tail fins for the control surface, and it would start rolling uncontrollably on launch when I activated SAS, making the vehicle impossible to control. Disabling the roll control on the tail fins instantly corrected this problem. 

[Archgeek]

11 hours ago, Workable Goblin said:

I disagree, because I've seen the same thing in rockets without any boosters at all. I built a simple single-stage rocket using tail fins for the control surface, and it would start rolling uncontrollably on launch when I activated SAS, making the vehicle impossible to control. Disabling the roll control on the tail fins instantly corrected this problem. 

Not trim, then.  'Sounds like your SAS input must be zero sans fins...hmm, do you use angle snap in the VAB?  If not, it's possible the editor's derping slightly and they're going on a bit twisted or the control surfaces........

Hmm.  Where on your rockets are you placing the fins?  A time ago there was a problem with control surfaces in front of CoM getting reversed, deflecting the wrong way for a given input.  Under SAS, that problem would definitely self-worsen as corrections were attempted...but that would also break pitch and yaw.

[JohnnRico]

23 hours ago, Snark said:

Yes.  "Post a screenshot, please." 

Okay

http://i.imgur.com/JNwC1ha.png So this is with SAS on. Everything more or less fine.

http://i.imgur.com/EDTzpTx.png and this is with SAS off and the Trimming rested. The probe is spinning like crazy. ALL indicators at the lower left side oscillate quite fast. 

[Snark]

2 hours ago, JohnnRico said:

Okay

http://i.imgur.com/JNwC1ha.png So this is with SAS on. Everything more or less fine.

http://i.imgur.com/EDTzpTx.png and this is with SAS off and the Trimming rested. The probe is spinning like crazy. ALL indicators at the lower left side oscillate quite fast. 

Thanks!  That helps.

So, I don't know whether something else might be going on as well, but I'll say this much:  your probe has overwhelmingly, stupefyingly, insanely too much torque on it.  That reaction wheel is incredible overkill.  Even the little 0.625m one would be overkill for a small, light probe like that.  Ditch the reaction wheel and rely on just the probe core's own built-in reaction wheels.

Here's a quick experiment that doesn't require rebuilding or relaunching anything:  Try turning off the big reaction wheel (so that only the probe's torque is active).  Does your problem get better?

That doesn't really explain why you're getting flutter when SAS is turned off-- it ought to just be dead-stick.  Do you make use of any clipping on the ship?  What's that black thing poking out below where the solar panel is attached?

Edited January 11, 2016 by Snark