Trying to control a goddamn mess to eject the bane of my KSP-existence out of the solar system

[THX1138]

A long time ago I regrettably accepted a contract to eject a class E asteroid from the solar system. I've sent a total of four different vessels to the same asteroid. I cannot control the attitude of the craft/asteroid - no matter what I do, I have oscillations that increase in intensity until something explodes. I need to eject this asteroid to break even on the contract!

I've made the save file available via my Dropbox. In the screenshot below everything looks fine but in actually, it's just passing through prograde.

I've tried disabling reaction wheels on some of the vessels in case they're doing something stupid.

I've detached the final craft (the one with the greatest diameter) and attached it to the asteroid directly).

A bug has given all the craft except the final one full fuel. I've no idea how or why but I'm not upset about it; it makes up for the useless oxidiser I've been stuck with since NERVs stopped using it.

How can I control what I have without sending any more vessels? What I have should be just fine in my opinion.

 

 

Edited June 22, 2016 by THX1138

[Foxster]

I think if it's been hanging about being a pest for this long then I'd be very tempted to HyperEdit its cheeks outa there.  

[bewing]

Even if you get it under control, it's gonna take a million years to get it up to speed and outta there. So I agree with Foxster. I can't see a way to do it with your current setup without constant hand-adjustment of thrust levels for the entire trip.

[Tatonf]

You need to upload the folder, not only the save file.

[FyunchClick]

jeepers. With such an expansive ship and so many connections I'm not surprised you got oscillation. Your SASs are also probably fighting each other. If redesigning your mission is not an option, perhaps look at something like KAS to strengthen or reconfigure what you have.

 

 

[pincushionman]

I'd abandon those long arms you're using to direct the thrust around the asteroid; there's no effective way to stiffen them sufficiently for what you're trying. Use a long girder pole between the puller unit and the Klaw, and angle the thrusters to miss the asteroid instead.

And instead of all those tanks, centralize to a few big ones, then glomp an ISRU near the Klaw connection, and suck out the asteroid's own innards to power yourself along!

…but if you're already committed to what you've got (or just a little more), ditch everything except your stiffest puller unit on the front, and stick an ISRU (with all control disabled) on the back. Then good luck.

Edited June 23, 2016 by pincushionman

[Just Jim]

Question.... slightly off topic.  I see on the forward long armed assembly you have a whole bunch of NERV engines.  But also a whole lot of standard rocket fuel tanks with both liquid fuel and oxidizer.  Why?  NERV's run on liquid fuel alone, and all that extra oxidizer is not just a waste, but also a lot of extra weight you don't need.  I would replace those liq/oxy fuel tanks with liquid fuel only tanks. 

[THX1138]

I've noticed that the SAS is frequently applying yaw in the wrong direction. This is very frustrating.

[Just Jim]

When I moved my Emiko Asteroid, which weighed in at just over 3,800t, I used a much simpler approach. 
First I attached 5 RCS pods, as I called them, to both stabilize and spin the asteroid into position. 
I positioned them around and in front of the asteroid as best possible.

Then I attached a big engine pod in back, which had 8 thuds surrounding a Rhino.

Finally I added a drilling and fuel converter pod.

Once everything was in place it moved fine.  

The burn times were really long, and I stopped a couple times to refuel, but in the end I was able to get it where I wanted it. 

 

Edited June 23, 2016 by Just Jim
forgot a pic

[Waxing_Kibbous]

Maybe this will help?

 

[Snark]

One of the problems you can run into with an asteroid is that SAS can actually fight itself and make the problem worse.

Let's say you attach your rocket and start thrusting, but you're not perfectly centered on the asteroid's CoM so it start to yaw slightly to the left.  How would you fix this?

Well, since your rocket is pushing the asteroid from behind like an outboard motor on a boat, what it would need to do would be to point the nose of the rocket as far to the left as possible-- because thrusting the rocket to the left will rotate the asteroid to the right, like a rudder on a boat.

Except that your SAS doesn't know this!  All it sees is "Oh noes, I'm getting off course to the left!  I should steer to the right to correct for that!"  which is exactly precisely the wrong thing for it to do.  It's because it's simply not smart enough to understand that it needs to apply "boat rudder" steering in a situation like this.

So, one thing you can do:  Separate your torque from your thrust.

Have multiple units docked to the asteroid:  one to apply thrust, and a bunch to apply torque.

The thruster unit, which you line up as precisely as possible with the asteroid's CoM, latches on and then completely disables ALL FORMS OF TORQUE.  Turns off every single reaction wheel, including on probe cores and such.  Locks the gimbal on any engines that have it.  Also, you want this ship to be as short and squat as possible, so that the engine is pretty close to the claw that's holding the asteroid.

Since the thruster has no torque available at all, it will do nothing with SAS and is always thrusting straight at the asteroid.  Of course, you will inevitably have a slight error from the asteroid's true CoM, so it will generate a bit of torque that way... but at least SAS won't make the problem worse.

Then, you have your torque units.  These are much like what @Just Jim shows.  You can have as many of these as you want.  If you're using only reaction wheels, it doesn't matter where on the asteroid they are or how they're distributed.  If they're using RCS, then you want them uniformly distributed around the 'roid.  These craft have no throttle-based thrust at all and are responsible only for generating torque to rotate the asteroid.  It doesn't matter if they wobble madly when they're applying their torque, because your control-from-here point is your thruster ship, and their wobbles won't affect it.

This layout works pretty well, and can sustain fairly high thrust loads.

Another approach you can take is to pull the asteroid instead of pushing it.  Your ship is way out in front of the asteroid, with the rear-facing claw at the end of a long, long boom trailing behind.  The ship engines are angled slightly to the side so that their exhaust misses the asteroid.  It means you take a slight hit to your fuel efficiency due to cosine losses... but on the other hand, you can thrust as hard as you like and it's rock-solid stable.

[pincushionman]

15 minutes ago, Snark said:

...Then, you have your torque units.  These are much like what @Just Jim shows.  You can have as many of these as you want.  If you're using only reaction wheels, it doesn't matter where on the asteroid they are or how they're distributed.  If they're using RCS, then you want them uniformly distributed around the 'roid.  These craft have no throttle-based thrust at all and are responsible only for generating torque to rotate the asteroid.  It doesn't matter if they wobble madly when they're applying their torque, because your control-from-here point is your thruster ship, and their wobbles won't affect it...

I would imagine this would be a good place for vernors rather than mono RCS.

[Jarin]

42 minutes ago, Snark said:

Another approach you can take is to pull the asteroid instead of pushing it.  Your ship is way out in front of the asteroid, with the rear-facing claw at the end of a long, long boom trailing behind.  The ship engines are angled slightly to the side so that their exhaust misses the asteroid.  It means you take a slight hit to your fuel efficiency due to cosine losses... but on the other hand, you can thrust as hard as you like and it's rock-solid stable.

Judging from the image, that ship is already pulling instead of pushing. 

[THX1138]

7 hours ago, Just Jim said:

Question.... slightly off topic.  I see on the forward long armed assembly you have a whole bunch of NERV engines.  But also a whole lot of standard rocket fuel tanks with both liquid fuel and oxidizer.  Why?  NERV's run on liquid fuel alone, and all that extra oxidizer is not just a waste, but also a lot of extra weight you don't need.  I would replace those liq/oxy fuel tanks with liquid fuel only tanks. 

Because when those ships were launched the NERV required oxidiser. NERVs only started to run on LF alone after a particular update which occurred after all these were launched.

[Gojira1000]

I've only ever pushed the things, using the gimble on the claw to get my CoM/CoT lined up, then hi-ho, burns.

[THX1138]

On 22/06/2016 at 10:51 PM, FyunchClick said:

jeepers. With such an expansive ship and so many connections I'm not surprised you got oscillation. Your SASs are also probably fighting each other. If redesigning your mission is not an option, perhaps look at something like KAS to strengthen or reconfigure what you have.

 

 

What situation results in SAS fighting itself?

[cantab]

On 23/06/2016 at 5:47 AM, Snark said:

Since the thruster has no torque available at all, it will do nothing with SAS and is always thrusting straight at the asteroid.  Of course, you will inevitably have a slight error from the asteroid's true CoM, so it will generate a bit of torque that way... but at least SAS won't make the problem worse.

I will add that if you have a cluster of engines, you can null out that offset by adjusting thrust limiters on individual engines. KER will show you "thrust torque" and you can tweak until this is small, or you can just do a bit of a burn with SAS turned off and see which way the asteroid turns and reduce thrust on the 'outside' of the turn.

On a smaller scale, such an engine cluster approach lets me easily move my lopsided station around, https://flic.kr/p/HCJqqS

 

Edited June 24, 2016 by cantab

[FyunchClick]

On ‎24‎-‎6‎-‎2016 at 2:43 PM, THX1138 said:

What situation results in SAS fighting itself?

SAS on the ends of noodly structures tend to start oscillating rather than stabilizing. The correctional inputs come from wherever you're controlling from rather than where the unit is located so that also doesn't help in getting stuff under control. In general you want your SAS connected with as little flex as possible to your control module, which usually means just a few connection links away.

I'm no 'roid wrangler, but if I got to design this from scratch I'd go with @JustJims suggestions. A rather stubby high impulse high power gimballed pushing engine, and lots of Vernors to augment control. My experience with pulling asteroids is that it's not particularly stable per se (rather the opposite). The idea that once the assembly comes under power it straightens out is false, it's easy to make the analogy with pulling a heavy boulder on a rope but ground drag will always point exactly opposite to the pulling force which naturally stabilizes, but such a force doesn't work on the asteroid and it'll just whiptail all over the place.

I particularly like the suggestion of planting RCS clusters around the asteroid as he suggested,

Edited June 25, 2016 by FyunchClick
spelling

[jedensuscg]

On 6/22/2016 at 9:28 AM, Foxster said:

I think if it's been hanging about being a pest for this long then I'd be very tempted to HyperEdit its cheeks outa there.  

You can also ctrl-12 to open up the debug menu, go to contracts tab and hit the complete button. This will treat the contract as completed despite the asteroid being still being where it was, give you the money and reputation and let you continue to experiment with said asteroid. 

[Diche Bach]

Capturing asteroids. So fricking cool. Can you like mine those things for platinum and molybdenum and excrements?