Experimenting with disabling SAS for the duration of the launch

[MedwedianPresident]

I've recently noticed that SAS, if activated prior to launch, does more bad than good when it comes to big, wobbling rockets: the counter-forces that the SAS enacts on a naturally wobbling rocket usually encache the oscillations instead of dampening them. Besides that, SAS does not do any good when it comes to keeping a certain orientation when the rocket is big and you will manually have to control the rotation anyway. For my current mission (an orbital bombardment system that requires a medium-sized launcher) I am currently experimenting with turning on the SAS only after the first stage has been detached, e.g. flying SAS-less during the first part of the ascent. The rocket still explodes when I start the gravity turn, but I have noticed that the wobble oscillations have decreased.

[Pecan]

12 minutes ago, MedwedianPresident said:

...The rocket still explodes when I start the gravity turn, but I have noticed that the wobble oscillations have decreased.

Sounds like progress.

[MedwedianPresident]

6 minutes ago, Pecan said:

Sounds like progress.

Now, the rocket explodes not at gravity turn but when I try to ditch the first stage; however, this has something to do with the interstage fairings and I think that I have fixed it now.

[GoSlash27]

This is probably a good way to go, especially when you're still new at KSP.

Turning off SAS eventually leads to gravity turning. Not only is it more efficient, but it's also easier once you get the hang of it.

Best,
-Slashy

[MedwedianPresident]

SAS actually does not help but instead hinders during launches. It is simply not made for larger vessels. It turns off automatically when a certain rotational velocity has been reached and doesn't help one to stabilize a spinning rocket. If one goes without SAS until ignition of the second stage and actually takes care of the rocket's rotation himself, the launch is safer and easier.

 

However, as you jettison more and more stages and your vessel becomes smaller and more stable, SAS becomes very useful and essential. 

[polytechnique]

It kind of depends on the craft for me ... when I'm making huge, STS-style shuttles using the vector engines, I have to use SAS the whole way up. Though, the structure does lend itself pretty well to adding a lot of struts, so it holds together plenty well.

Just depends on your construction I suppose...

[ImmaStegosaurus!]

How many engines you use, which engines and have you limited gimballing on those engines?

9/10 times my rockets oscillate it's solvable by tweaking gimballing.

[MedwedianPresident]

The first stage is a cluster of 7 SRB's.

[Hodari]

6 hours ago, ImmaStegosaurus! said:

How many engines you use, which engines and have you limited gimballing on those engines?

9/10 times my rockets oscillate it's solvable by tweaking gimballing.

Gimballing or more struts.  But yeah, if it's oscillating that much then there's something wrong with the design and not just a problem with SAS(though SAS certainly has its own issues as well).

[VincentThacker]

I don't think all big rockets are negatively affected by SAS. My Space Shuttle stands at almost 450 tons (475 with maximum payload), with lots of struts connecting the solid boosters, orbiter, and external tank, and does very well with SAS. At first, (when I was very noob), I did not use struts, and the solid boosters wobbled, and the whole thing fell apart. With struts, and SAS, it is very smooth. You've just got to strut it more (and at the right places).

By right, SAS is supposed to stabilize by exerting opposite forces to the current direction you're rotating towards. The reason why it does not do so on larger rockets is because they "bend" too much. When a rocket "bends" (oscillates) too much, by the time SAS corrects the pitch/yaw/roll of your rocket, your rocket will have oscillated to the "other" side, making it even more unstable. In other words, your rocket "bends" faster than SAS can correct each oscillation. That is the reason why "bendy" rockets tend to  break when using SAS.  If you find it hard to control your rocket, try making your rocket less "bendy" (make it more rigid). Also try to add control surfaces or more vectoring engines and see if it helps.

Edited August 8, 2016 by 100055

[Mister Kerman]

Another good trick. Disable or reduce the gimbal of radial engines. Too much gimbal can cause over-correcting.

Usually if your craft isn't a strand of spaghetti, just the gimbal from your main lifting engine ought to be enough to keep you pointed straight.

Edited August 8, 2016 by Mister Kerman

[Garrett Kerman]

Really? So THATS why my craft wobbles. I thought SAS would help. Well then... (Turns off SAS)

[MedwedianPresident]

Another trick: if not even disabling SAS helps, your rocket may be too small for the payload. I tried to lift a 80 ton sattelite using a medium-size launcher but it turned out that the only appropriate rocket for that was a giant monster made out of 7.5m Space-Y parts.

[samstarman5]

For large rockets, it really makes a difference where you place your reaction wheels.  If placed on the command stage, I may very well have problems during launching(Although not always). However, if I move it a stage down, closer to the center of mass and thrust, there is less wobble.  I find if I have a small command stage then I don't need a reaction wheel there as what the pod has does enough when it is down to that stage.  For larger then I would include one, but leave it off to allow the launch stage's stabilization to control.