Aligning The AGU With A Target's CoM?

[The Flying Kerbal]

Hey guys, I'm stuck on this one...

Since I unlocked the Advanced Grabbing Unit, I've been getting contracts to not only rescue stranded Kerbanauts, but to also return the wrecks back to the surface of Kerbin.

After a few initial experiments I've found that hitting the target at 5 m/s seems to guarantee a successful grab, but I've no idea how to align the grabber with the target's CoM.  So far I've only accepted one of these contracts (by accident!), and had to return a heap from a Munar orbit back to Kerbin.  I did manage to do this, but it was amazingly difficult with my transfer burn back to Kerbin being done at less than a third throttle to have anything resembling directional control of the spaceship.  It was lucky I left Kerbin with way too much fuel for a normal run to Mun, because with all the fighting and twisting to stay on course, I was running on fumes by the time I got the Pe into the planet's atmosphere.

I've looked on Youtube for videos, and on this and other forums for threads concerning this, and while there is some information, I must be honest and admit that this time they didn't really help very much.  Maybe they were more aimed at asteroid capture, but certainly they didn't do much for me struggling with bits of spaceships.

So how do you guys line your AGU's up with a target vessel's CoM?  I have a contract waiting to be fulfilled to return an A Type asteroid to a Kerbin orbit (this will be my first attempt at this), and I do think there is some sort of aid for targeting the CoM's of asteroids, but really if it was as big a struggle as I've had so far, it really wouldn't be worth the effort.

Any help, hints, tips or advice will be very much appreciated, oh and just as a little incentive, another Kerbal has managed to get himself stranded in a Munar orbit and Mission Control wants his vessel returned too.  If you don't want to help me... just think of that poor Kerbal and his family... 

[bewing]

You can't see CoMs in flight, except for asteroids. So either you guess, or you make sure to touch your target along its central axis, or you do your remaining burns at very low thrust so that your SAS, gimbaling, and reaction wheels can handle the CoM offset.

2 hours ago, The Flying Kerbal said:

I've found that hitting the target at 5 m/s seems to guarantee a successful grab

Pro tip: it has nothing to do with speed. To grab something, your klaw has to be perpendicular to the surface it touches. This is easiest if you are trying to grab a flat face. On a curved surface, you also have to aim for the centerline of the part.

 

Edited November 23, 2017 by bewing

[Helmetman]

When using the AGU I put it on a longer arm made of structural parts or LF tanks and then attach the engines to the far end. Angle them a bit so they don't fire against the module I'm willing to transport. Then I pull the craft rather then pushing it.

If you push it like you would normally as most of all rockets are like that you will have a greater balancing issue over towing. It is not recommended on rocket in a atmosphere by the way, just to let you know.

Try to push a box by standing tall to the highest reachable window. You'll use 2 hands blimey and you still whining how it almost fell out of your hands. Rather, let someone grab the box on the next floor who tows it from your hands over the railing in a second or 2. In part this is because it's harder to push things above your head, but like I said, in part.

Pulling it causes the CoM to trail towards the thrust vector.

This means you have to do less of correction because it will auto correct itself, if you have to correct at all. Pushing it causes the CoM to trail of off the thrust vector, like a stick will fall of your hand trying to raise and balance it. This makes a stick or your vessel easily turn over when firing to hard. Pulling it means placing the thrust vector in front of what your willing to transport like a crane. In that case, the stick of the guy below would be hanging in it's hand, rather then balancing it. 
Because that's what you do when pushing, your basically balancing a candle. 

Basically it's this principle 

vs this principle 

 

If you do go for pushing then makes sure your engines are farthest away from the CoM (you can eyeball this)
This means the relative offset of CoT and CoM becomes less. And only if you have sufficient gimballing range on your engines ( I recommend the vector ) the distance from the CoT to CoM creates a longer levering arm. That means that less pronounced control strokes will give similar corrections. Of course you don't want it to long otherwise it will become wiggly especially under high thrust.

I hope this has helped you.

2 hours ago, The Flying Kerbal said:

  If you don't want to help me... just think of that poor Kerbal and his family... 

Buahaha, getting our attention by exchanging sentiment, Great. I hope it was only your humour since we're talking pixels, but yeah very poor indeed  

[Rocket In My Pocket]

As explained above by @Helmetman pulling is always better than pushing when it comes to moving objects in space.

As an alternative method, I'll often put the Claw near my craft's COM as opposed to way out on one end, so at least I know I'll be carrying the weight were it will bother me the least. This method works especially well when the craft is significantly heavier than the target.

This space shuttle for instance, carriers it's Claw in it's payload bay. (Ignore the fuel tank/debris floating away, that's extra fuel I bring up in the empty payload space, it get's jettisoned before bringing the target in; the Command Pod in this case.)

Edited November 23, 2017 by Rocket In My Pocket

[Kryxal]

I'll use 4 engines, with radially-attached stacks.  This means I can adjust their thrusts individually for fine-tuning.  If I can get it right, then throttle's easy.

[Pand5461]

5 hours ago, Helmetman said:

Because that's what you do when pushing, your basically balancing a candle. 

 

2 hours ago, Rocket In My Pocket said:

As explained above by @Helmetman pulling is always better than pushing when it comes to moving objects in space.

Guys, I'm sorry to contradict, but you're falling into the "pendulum rocket fallacy" here. In short: there is no difference between pushing and pulling if your engine is rigidly connected to the weight it moves.

@The Flying Kerbal AFAIK, target marker on the navball points to the CoM of vessel or asteroid. So, if the Klaw's axis is on one line with engine thrust vector and "grabber" ship's CoM, then aligning both relative velocity with target and active ship orientation with the target marker should do the trick.

Another way is using the pendulum effect properly, i.e. when connection with the grabbed object is flexible. To make it so, click RMB on the Klaw once it's grabbed the target and release pivot. Then, with released pivot do an RCS burn backwards (if Klaw faces forwards) or forwards (if the Klaw faces backwards). This can correct small misalignments of target CoM from the thrust vector. After the correction, lock the pivot again and continue your mission.

[Rocket In My Pocket]

9 minutes ago, Pand5461 said:

Guys, I'm sorry to contradict, but you're falling into the "pendulum rocket fallacy" here. In short: there is no difference between pushing and pulling if your engine is rigidly connected to the weight it moves.

Another way is using the pendulum effect properly, i.e. when connection with the grabbed object is flexible. 

Nothing is ever connected rigidly in KSP, so pushing a large mass from behind almost always results in bending.

We understand pendulum fallacy, but this is fairly well established "KSP logic" we are talking about here. You want to pull an asteroid (Or any large mass), not push it.

This comes down to how the physics system works in the game, it's fairly bendy by nature. Even with autostruts and rigid attachment, it's fairly impossible to make a 100% rigid craft; especially once you've docked or grappled it to another craft/asteroid. These "in-situ" connections are even more "wobbly."

Edited November 23, 2017 by Rocket In My Pocket

[Helmetman]

@Pand5461 If only the guy could glue the stick to his hand

But realistically (non ksp'ish) your very right  

[Pand5461]

13 minutes ago, Rocket In My Pocket said:

This comes down to how the physics system works in the game, it's fairly bendy by nature.

Pretty rigid to me, ever since autostruts were introduced. And I usually use a wide ship rather than a long one when I need to move an asteroid.

[Rocket In My Pocket]

4 minutes ago, Pand5461 said:

Pretty rigid to me, ever since autostruts were introduced. And I usually use a wide ship rather than a long one when I need to move an asteroid.

That works, more than one way to skin a cat.

Surely you can agree though that if at worst there is no difference, and at best it alleviates the wobble in your ship; it makes sense to pull rather than push. It certainly won't hurt you or have any negative effect so why not?

Considering wide is worse for launching, I'd wager the average player will end up trying to move an asteroid with a pretty standard looking rocket; long and thin...and bendy. In this case, they should definitely pull.

 

[OHara]

You can adjust the alignment after your are connected (just being sure you know) using "control from here" on the claw, "target center of mass" on the asteroid,  and "free pivot/lock pivot" on the claw.

 If you push, and use engine-gimbal to steer, that does tend to flex the joint in the claw.  The trouble starts when the flex turns the pusher more angle than the gimbal-angle of the engine.   Short pushers make things easy, giving less lever arm to flex the joint, but still plenty of lever arm to turn the target.

 

Edit: Oh yeah, you said non-asteroid targets.  You can free-pivot the target, use a tiny puff of RCS to reverse, and that starts the target's COM swinging toward the line of thrust of the RCS.  A few cycles of free-pivot / back-up / lock-pivot, letting the target swing only a small angle before locking the pivot, will leave the target settled into position along the line of thrust ... of the RCS, which is usually the same as the line of thrust of the main engine.

Edited November 24, 2017 by OHara

[The Flying Kerbal]

Great answers as usual here guys, thanks very very much!

I sent a Klaw up to Mun to rescue Chadfal Kerman and bring his vessel back to Kerbin.  This time I tried the technique of unlocking the pivot and thrusting backwards, and it seems to have worked, I wonder what I did wrong before?

Another little issue cropped up as the recovery package was coming back into Kerbin's atmosphere.  Rightly or wrongly, I have the AGU set up to separate from the command module "lifeboat" the guy being rescued returns home in.  I'm sure you guys would have a much more elegant way of doing this, but I know the Mk 1 capsule will have no problems during reentry, whereas the AGU with the wreck attached can't be trusted and needs watching the whole way in.  So the idea is to let the command pod reentry by itself, the chutes set to open when things were nice and safe.  Both sections are landing on water, so nothing's hitting mountains or meeting an untimely demise in some equally unsavoury manner.  Yet I can only save whichever vehicle I'm actually watching, the other seems to just vanish?  I know debris is automatically removed, but these are both being identified as "Ships", so that doesn't seem to be the problem.  After hitting F9 many times, I finally had to pull the Pe of the Mk1 Command Pod out of the atmosphere, and send it round the orbit for a second go, lucky I had some fuel left in the tank.

Does anyone know why this is happening and how I can get round it?

Thanks again everyone!

[bewing]

Unless two craft are in physics range of each other (about 25km) -- there is an important concept called the "autodeletion altitude". There is only one physics bubble in the game, and it's around the ship that you have focused. Only ships within that physics bubble are modeled. Any other craft is just stored as a placeholder and is just falling with gravity. If an unfocused ship goes below the autodeletion altitude, it gets autodeleted. (What a concept!). The autodeletion altitude on Kerbin is 25km. It doesn't matter if it's a kerbal or a ship or a probe or an airplane. Anything that is outside your physics bubble on Kerbin, that is flying below 25km altitude is assumed to have crashed into the surface, and is deleted.

[The Flying Kerbal]

1 hour ago, bewing said:

Unless two craft are in physics range of each other (about 25km) -- there is an important concept called the "autodeletion altitude". 

I didn't know this, I always thought if it had an active core or pilot, then it would be allowed to land safely on its parachutes.

That'll certainly save me a lot of head scratching trying to figure out how to land both vessels at the same time.  Thanks for answering.

[bewing]

Nope, parachutes do not work on craft that are outside the physics bubble. Neither does Aero drag. A craft outside the physics bubble will just fall and impact the surface at full orbital velocity.

 

[Zeiss Ikon]

On 11/24/2017 at 6:10 AM, bewing said:

Nope, parachutes do not work on craft that are outside the physics bubble. Neither does Aero drag. A craft outside the physics bubble will just fall and impact the surface at full orbital velocity.

And the take-away from this is, if you have multiple craft that you need to land, either keep them close together (which I've done with command pods that I decoupled just as reentry heating started in earnest) or leave them in a stable orbit, except for the one you're actually landing.  For career, this is important because an auto-deleted vessel can't be recovered for your partial parts cost refund.