Problems with SAS/ASAS

[Fizwalker]

I don't know if anyone has had this happen, but as of my last two launches, it feels like ASAS/SAS isn't there. The ship pulls to one side, spins randomly over 2 axis. This effect occurs outside of atmosphere as well as in it. Engines on or off. Turning off the torque along with SAS/ASAS seems to mitigate it somewhat. So far, the only solution I have found is closing the client and reloading it, but that is only a temporary solution as it returns after playing for awhile.

Thank you for your help in advance.

Edited June 26, 2014 by Fizwalker

[Greenfire32]

Is this happening on a multi-craft (2 or more docked crafts)? It sounds like multiple ASAS/SAS are fighting with each other over who gets to "control" the craft. If this is the case, the solution is as simple as disabling the child units, while keeping the parent unit online.

[Taki117]

This sounds like there isn't enough control authority. A picture of a problem craft would help, but depending on rocket size, pod torque may not be enough and you need to add more reaction wheels.

[cantab]

Do you have engines forward of the centre of mass, for example in "puller" arrangements? SAS assumes engines are aft of the CoM and will gimbal them the wrong way when they're forward. Solution is to lock the gimbals on the engines and rely on other means to steer.

[Kasuha]

Picture of the ship would definitely help.

Reaction wheels cannot fight each other, they always act in the same direction. But your ship may be too wobbly and may be getting into resonance. Or there may be delay in transferring momentum from your reaction wheels to your control module.

Or you may perhaps have some mod installed that gets confused and is causing all this trouble.

[Epthelyn]

Are your controls locking up somehow (as in the game thinks a key is held down when it's not)?

If the problem is resolved by reducing the torque on the rocket then it sounds like either that or the SAS playing up, but since it takes a while to happen during a play session I'd have thought any problems with SAS misbehaving that badly would show up immediately, not "a while" later.

That is if the rocket is a simple single-'part' (no docking etc.) craft, anyway.

As Kasuha said, pictures of the ship would help.

[Fizwalker]

Thank you all for your responses, it's much appreciated!

I should clarify a little. Launching the ship was a caste iron *****, and once in orbit it just got worse after shutting down the engines. I tried quicksaving and reloading that, with no effect after trying to control it. After that, I shut down the client to desktop, and re-launched the game. After reloading the client, and save game, the issue disappeared, though at this point both vessels were in orbit at that point. With the first ship, the problem resurfaced after reaching Minmus orbit.I'll have to post pics of both ships in question (Two separate launch vehicles, with no docking operations.) probably tomorrow though.

Note: I removed the fairing for clarity, and added the CoM marker.

This was the vessel's build upon entering Minmus's SOI, and the problem's re-occurrence. (I was in the map, working on the orbital insertion maneuver when I saw the Navball start spinning wildly.

The second ship in question, I didn't save the booster with the payload unfortunately, and I haven't tried launching it since (upon reloading the client in this instance and selecting the ship to fly, it violently disassembled itself, killing Jeb and Bill.) However, I can post a pic of the payload if that would help.

Again, thank you all for the responses.

Edit: Pictures didn't show up when I posted it first....

Edited June 18, 2014 by Fizwalker

[Kasuha]

It seems to me the main source of torque is the large reaction wheel at the top of ascent stage and there is a rather wobbly connection between it and your control point.

Possible solutions:

1/ use struts around the small decoupler to make its connection sturdier, also add struts bethween the ascent part and the upper module

2/ add a large probe core right next to the large reaction wheel and control the ascent from there all the way until you decouple that part. Only after that switch to the probe core on your payload

[Kryxal]

Solution 2 also has the benefit of being able to use stage 2 clear to orbit, then turn around and de-orbit it after decoupling.

[stibbons]

Did you mean that you're seeing torque on your vessel even with engines and SAS off? In that case, make sure that you don't have any parts clipping through each other.

[Fizwalker]

Again, thank you for the replies. I have rebuilt the booster for the 2nd ship that has demonstrated the problem. I have no clue as to where the problem is originating from at this point. As you'll see in the pics of it, the only SAS/ASAS is from the PPD-12 Cupola module. In this case, launch was a little squirrelly but manageable. Once in orbit, I ditched the final stage with an engine on it, and the station piece proceeded to start tumbling. I first turned off the torque in the Cupola, and the rate of acceleration in the tumble slowed for a bit, but did not stop. I then turned off SAS completely. Again, it slowed down somewhat for a bit, but then picked up.

Stibbons mentioned part clipping, and I went through the design again, and noted that 4 of the RCS thrusters were clipped inside the Cupola as well as one of my comms antennas. I replaced the RCS thrusters with stock versions, and shifted the antenna with no effect on the outcome. (The pics are of the station piece prior to the clipping removal alteration.)

1st from inside the VAB (again, fairing removed for clarity)

2nd is of the insertion stage and payload.

http://www./convkey/df4e/q5plo08a0xtdq6afg.jpg

3rd is just the payload in orbit.

http://www./convkey/b57e/yiaqhxdkdvxf0emfg.jpg

Again, thank you all for your help and suggestions!

[Kasuha]

It is not clear to me what do you call 'tumbling'. Your orbital stage is not going to be at rest relative to Kerbin - it is at rest relative to universe. If you point your Cupola module at Sun, it should stay pointing at the Sun. But on night side when Kerbin comes in the way it will be pointing at Sun through Kerbin. And when on day side it will still keep pointing at the Sun, away from Kerbin.

[Fizwalker]

I wouldn't call what you describe as tumbling, as the ship is stationary relative to its axis, using your screen and the ship as a point of reference. Yes, it has a really slow rotation in reference to Kerbin, but just watching the vessel, it doesn't appear to move. I'm not sure if that makes sense. I'm not really sure how to explain that clearly.

In regards to my station, perhaps this will help? I created a gallery on imgur and posted a number of screen shots from one rotation of the ship in question. They are in sequence. Keep in mind that the total time of the rotation was under 15 seconds or so.

http://fizwalker.imgur.com

(There's only 1 album so far--The Wonderous Tumbling Station--and I'm sure there's a way to embed it here--I've seen it in other posts!--but I don't know how to do that.... >.< )

Edited June 18, 2014 by Fizwalker

[Claw]

You embed imgur albums by using the imgur tags plus the five letter identifier from the imgur URL.

(imgur)5-ltr(/imgur) except use the square brackets...

Okay. Try this...

Turn off all torque (not just SAS, but disable the torque), engines, RCS, SAS, and press ALT-X.

Tap the time warp key and all rotation should stop. Then go back to 1x warp.

If the ship starts to tumble (watch the navball), then you've got either a design problem (parts clipping) or a mod that is interfering.

If it doesn't turn/tumble, then start activating things one at a time.

Good luck,

-Claw

[Fizwalker]

You embed imgur albums by using the imgur tags plus the five letter identifier from the imgur URL.

(imgur)5-ltr(/imgur) except use the square brackets...

Okay. Try this...

Turn off all torque (not just SAS, but disable the torque), engines, RCS, SAS, and press ALT-X.

Tap the time warp key and all rotation should stop. Then go back to 1x warp.

If the ship starts to tumble (watch the navball), then you've got either a design problem (parts clipping) or a mod that is interfering.

If it doesn't turn/tumble, then start activating things one at a time.

Good luck,

-Claw

Quick question Claw, what does Alt-X do?

[Claw]

Quick question Claw, what does Alt-X do?

Resets the trim. With SAS off, you can see if there is trim input with the indicators in the bottom left. ALT-X sets them back to the middle.

[Fizwalker]

<iframe class="imgur-album" width="100%" height="550" frameborder="0" src="http://imgur.com/a/cWgDK/embed"></iframe>

Testing this

[Specialist290]

For the record, the HTML embed codes won't work here. What you can do, though, is take that little identifier at the end of the url and put it inside IMGUR tags, so that this (minus the asterisks):

[imgur*]cWgDK[/imgur*]

Becomes this:

Javascript is disabled. View full album

[Fizwalker]

Resets the trim. With SAS off, you can see if there is trim input with the indicators in the bottom left. ALT-X sets them back to the middle.

Between you and specialist, you've shown me the way, thank you very much!

Damn. thank you Sir!

Javascript is disabled. View full album

This is a test of the imgur album syntax

Edited June 19, 2014 by Specialist290
Merging sequential posts by same user. Please make use of the edit function if no one else has posted in the meantime.

[Claw]

Although I suppose to be complete, I should have also included to check joystick settings.

Regardless, I take it your problem is all fixed?

If so, your last lesson is how to update the status of your thread. Edit your first post and click "Go Advanced". You can change the thread prefix via the dropdown, then save.

Edited June 19, 2014 by Claw

[Kasuha]

[noparse]

Javascript is disabled. View full album

[/noparse]

This is a test of the imgur album syntax

You need only to use the imgur hash without http address and image index for that to work correctly, i.e. not

[noparse]

Javascript is disabled. View full album

[/noparse]

but

[noparse]

Javascript is disabled. View full album

[/noparse]

And the forums also support the [nop[noparse]arse]...[/nopa[/noparse]rse] tag, it is useful when you want to show how to make other tags.

[Fizwalker]

Awesome, thank you for the help!

I am beginning to suspect one of the mods I have installed might be the issue. I removed all the parts from a specific mod off that station piece, and launched it. The launch went smoothly, it didn't pull to the south like the previous launches. Once in orbit, it maintained attitude with the only rotation being that of a normal orbit. I.e If I pointed the cupola at the sun, the cupola stayed pointed at the sun, regardless of the station's position in it's orbit. The problem seems to be with AIES, but, not with all of the parts as I have launched other rockets using that part pack without problems previously. So I suspect that I have more testing to be sure of it, or even if it's just in certain positions that those parts have issues with. (I noticed that the RCS thrusters and RTGs were mostly inside the side of the cupola. I also noticed that the antenna's base above the crew hatch had clipped inside the roof of the cupola.)

So at this point, I believe the question is tentatively answered. I'm going to kick it around a bit more before I switch the prefix of the thread to answered. Again, thank you all for the responses and the help!

[Kasuha]

I also noticed that the antenna's base above the crew hatch had clipped inside the roof of the cupola.

Many people think that part clipping negatively affects ship's physical characteristics but so far I have not found any evidence for it.

Ship's parts cannot collide with each other. If you make a ship with long parallel arms and get it into motion you can notice its arms can freely pass through each other. There are only two ways how clipped parts can make things worse:

1/ when there is a lot of parts clipped into the same place, they make that point of the ship very heavy which might lead to increased strain on connections nearby when some force is affecting the ship (usually gravity, thrust, or torque).

2/ when such clipped part gets detached from the ship (usually because its connection to the ship fails) it suddenly starts checking for collisions with parts of its former ship and since it overlaps with them by large part of its volume it usually explodes violently and starts a chain reaction where more and more clipped parts get detached and also explode because of overlaps.

If you have one part that partially overlaps with another part or two, there's usually no problem at all.

[Claw]

Many people think that part clipping negatively affects ship's physical characteristics but so far I have not found any evidence for it.

...

If you have one part that partially overlaps with another part or two, there's usually no problem at all.

To piggyback on this... I typically find that part clipping allowed by the editor isn't a problem. It seems to do a reasonably good job of preventing too many things from overlapping.

It's not usually a problem till you enable the debug ALT+F12 clipping and start stacking things up (like Kasuha said). If you aren't careful there, that's when it'll cause all kinds of problems. I've seen parts jimmy lose doing this and the rest just comes apart.

[Fizwalker]

That's usually the case too for me. But of recent times, I've found ships violently disassembling themselves (not to be confused with exploding, but all the parts speeding away from each other at high velocity, usually with high rotations about individual CoMs per second.) Which I suppose is related to the initial problem in that they occur after reloading the game from desktop and then jumping to said ship from the Tracking Center.

(I don't ever use that ALT-F12 tool)