Rockets start rotating at launch

[Magzimum]

I started playing 1.2 since it came out. In my new career, I have now reached the point that I build (slightly) larger rockets that also have SRBs and/or additional LFO rockets on the sides. I notice that very often these start to rotate along the vertical axis when I start the gravity turn. I used the symmetry tool in the construction, so it is certainly not a lack of symmetry in the design.

From an aerodynamics point of view, it makes sense to me: the rocket is (for a short while) not pointed exactly prograde, so the booster that is on the outer side gets more drag. If that booster is turned even the slightest to one side, that drag will start to rotate the rocket. This keeps on going as you make corrections. I would not be surprised if real rockets also experience this - although most rockets don't use nearly as many radial attachments as we use in KSP. 

But from a gameplay point of view, I just want to get rid of that rotation, because it screws up a LOT of my launches... Any tips? 

I made an Imgur album with 3 examples of rockets that show unwanted rotation (below, in spoiler), but please note that this is a general question, not one about these examples. 

Spoiler

Small ship - rotates a little, and only when a low trajectory is chosen. The ship uses only SRBs here, with fins to steer. Quite easy to fly. 

 

Medium ship. Rotates very noticeably. Requires constant correction (pressing Q and E). Can make it to orbit, but not if you look the other way for 2-3 seconds. 

 

Large ship. Rotates as much as the medium ship. Requires constant correction. 

 

Edited October 18, 2016 by Magzimum
fixing imgur stuff

[Clipperride]

I've also noticed that rockets tend to rotate more since the 1.2 update.  One thing I spotted about your designs is, they don't have any tail fins.  Not only can that help keep the stack more stable, if you use fins that double up as control surfaces, you have more control in the dense lower atmosphere.  Hope that helps

[Aegolius13]

I've noticed the same.  You could try MOAR REACTION WHEELS, but I'm skeptical that will help.  I seem to be able to manually correct the rotation, but my SAS will not do it automatically.

[Tyko]

I've noticed my rocket tends to rotate a few degrees too..in 1.1.x builds if the rocket was balanced it would tend to stay on the 90 degree heading. With 1.2 it tends to rotate to 88 degrees or so. Then it's a fun game of trying to get it back on course  

[Hodari]

Also make sure any radially attached boosters have enough struts to hold them in place.  A lot of times for me, that sort of rotation seems to be because the boosters' thrust is pushing them out of alignment.

[Streetwind]

The reason for this behavior is the change in SAS behavior.

In order to facilitate the much more intelligent and smooth SAS position approach, the "aggressiveness" with which the system clings to an existing heading, or approaches a nearby heading, had to be lowered. So if you're already pointing where you want to be going, SAS will apply only very little stabilizing/correcting force. As such, it will not counter construction-induced roll nearly as well as it did in the past.

[Magzimum]

Thanks for all the feedback (I am getting used to the quality of the answers here, but I don't think we should take the quality of the forum for granted). 

What I learned from you all is that the rockets indeed behave a little different in 1.2. I will have to unlearn some behavior (in design, and in flight). I think that with more time, I will probably get a feeling for it. The designs I have here may be a little too draggy, and could probably be strutted down a bit more tight. At the same time, they are making it into orbit, so I am not getting stuck.

A little feedback from my side: 

• Tail fins: The first design has some tail fins. Also, after reading the feedback, I equipped the "medium size" design with tail fins (8 in total, one on each radially attached part), and the rotation did not reduce. While tail fins help if you have a rocket that is draggy at the top, this was not the solution for me. 
• The radially attached parts are strutted, but typically only with one strut at the top of the part(s). Movement of these parts relative to the central tanks is minimal (not visible during flight), but I cannot exclude that this cannot be made better. I think that "Moar Struts" may be a solution here. 
• I am skeptical that more reaction wheels will help so early in the launch. I feel that the tail fins and especially a Mainsail engine (gimbal at 100%) can correct the heading much better, and the input of the reaction wheels will be lost among that violence. 

Btw, in my career game, the problem is solving itself. I will soon have the 3.75 parts, and can then launch much larger payloads using just a single tank at the base. If I need to expand even wider than that, I'll come back to this thread!  

[Sharpy]

Symmetry or not, two levels "out" things do get bendy/wobbly. If you don't strut everything hard, the outermost boosters love to angle themselves a tiny bit within the allowance of the joints. That angle introduces a thrust offset - and as result, rotation, which will angle boosters on the other side as well. And quite stably, your outermost boosters remain angled from vertical, providing a corkscrew style rotation pattern.

[Harry Rhodan]

1 hour ago, Magzimum said:

 I equipped the "medium size" design with tail fins (8 in total, one on each radially attached part), and the rotation did not reduce.

Well of course. If your rocket is rotating because the radially attached parts bend in one direction, attaching fins to the bending parts won't be more stable. At worst that'll introduce even more torque.

Edited October 19, 2016 by Harry Rhodan

[Clipperride]

1 hour ago, Magzimum said:

Thanks for all the feedback (I am getting used to the quality of the answers here, but I don't think we should take the quality of the forum for granted). 

 

This is certainly the most informative and helpful game forum I've ever had the privilege of using.  From the simplest* question from new players to deep technical problems, there's someone here willing to help.

*And as "simplest" is a relative term, I really like the fact that newcomers aren't made to feel their questions aren't just as important as any one else's.  The forum is full of people that want others to enjoy the magic of KSP just as much as they do.

[Cerberus738]

I'm having this problem too.  SAS is causing the roll.  Right at launch SAS just pins the torque one direction and its roll time. I can counter the roll manually and stabilize it and then SAS just pins it the other direction.  yaw and pitch work fine other than the constant rolling throws it off. Its only in the lower atmosphere.  Fins. no fins. gimbal. no gimbal. fairings. no fairings.  its frustrating.

[DD_bwest]

2 minutes ago, Cerberus738 said:

I'm having this problem too.  SAS is causing the roll.  Right at launch SAS just pins the torque one direction and its roll time. I can counter the roll manually and stabilize it and then SAS just pins it the other direction.  yaw and pitch work fine other than the constant rolling throws it off. Its only in the lower atmosphere.  Fins. no fins. gimbal. no gimbal. fairings. no fairings.  its frustrating.

the very first thing to always check, what color is your navball?

[bewing]

If your craft is stock then posting a craft file can help a lot, too.