[1.2.0.1] Stabilization computers interacts strangely with sky anchors [Resolved]

Saelem Black

Tinker, Tailor, Soldier, Spy
Joined
Aug 25, 2018
Messages
273
Likes
822
Points
205
#21
ZG, I think we've come to more or less a consensus, but do draw the free-body diagram at some point for yourself. (That's what I was doing last night, hah!) I think we're using different terms to describe the same thing. Fundamentally it's about keeping the gravity vector and the anchor vector pointed in exactly the same direction. Moving the anchor to the center of mass is one way to do it, but there are other ways, too, such as multiple anchor points or some method of pitch/roll control. As a result, the simplest fix for my airship is simply a third anchor. Stays up fine.

1560005871399.png

Now then, here's another question. You've shown that the stab computer will resist yaw. But will it resist pitch or roll? Especially given that the computer doesn't try to control lift (I really wish it did so we could actually hover with LMEs and rotors).
 
Likes: ZeroGravitas

ZeroGravitas

Breaker of Games
Joined
Jun 29, 2017
Messages
3,491
Likes
4,579
Points
675
Age
36
Location
UK
Website
www.youtube.com
#23
It does actually, but only with hoverbugged techs.
Well, the stab comp always limits velocity in all axes (3 linear and 3 rotational), to under 2mph (and not sure what angular velocity - hard to tell).

I didn't think hoverbug techs are special in this regard? The opposed pad pairs just have intrinsically strong hold against motion up/down with relation to themselves, regardless of the presence of a computer. Right?


ZG, I think we've come to more or less a consensus
Does that mean that we can (re)mark this thread as "[Resolved]" then? As there's no glitch/bug/code issue here.


Fundamentally it's about keeping the gravity vector and the anchor vector pointed in exactly the same direction.
But the gravity vector is always straight down towards the ground, while any small purturbation in the tech's position will offset the force from the anchor tether from being 180degrees opposed. It's all about finding stability *despite* disturbances to position and rotation of the tech, which requires negative feedback or some type.

I'd been considering the tech's lift vector (as more important), because that is changed by the roll/pitch angle of the tech and will increase the tech's horizontal offset.

Have (at least) 2 anchor tethers off-set by a good distance in one linear axis will obviously limit the ability of the tech to rotate in that axis, as the tether on the rising side will stretch and pull back harder while the falling side will slacken and reduce the pull, fighting the rotation. Hence why your tech never fell over forwards.

But this definitely isn't a universal solution because the anchors must be fairly far apart relative to the length of the tether, in order for them to have sufficient leverage. With the incoming doubling in the height limit of sky anchors (from 40 to 80m from ground level) this geometric stabalising effect will be reduced for your revised tech (although I'd expect it will probably still stay sufficiently stable).

The real problem is for smaller-modest sized sky techs that don't have enough width to use this trick. e.g. turrets, small bases, small airships/fliers, etc. Ensuring a mid-height anchor block position (inside of them) will still help these greatly.


Incidentally, I noticed (by testing with a free floating AG tech) that the forces exerted by the tethers are a little weird - they actually provided zero resistance up to 40m vertically, while immediately providing resistance to any horizontal offset:
Edit 2019-06-08: oh, I see - the tether has magical properties! It has a slack zone, exerting zero force for when extended up to 40m straight up above it. (Then a lineally ramping up force, like an elastic band.) But it immediately pulls back against any horizontal offset above zero. Weird, but practical, I guess.:)
I think that they probably do add a physically logical a downwards force component with their sideways forces, though. But tricky to be exactly sure.
 

Saelem Black

Tinker, Tailor, Soldier, Spy
Joined
Aug 25, 2018
Messages
273
Likes
822
Points
205
#24
Well, the stab comp always limits velocity in all axes (3 linear and 3 rotational), to under 2mph (and not sure what angular velocity - hard to tell).

I didn't think hoverbug techs are special in this regard? The opposed pad pairs just have intrinsically strong hold against motion up/down with relation to themselves, regardless of the presence of a computer. Right?
The stab computer doesn't directly control the vector setting on vectored-lift blocks, so it doesn't matter how close to the hover point you are, you'll eventually start accelerating in one way or another. I think the code reads the vector setting as "direct input", and so doesn't override it. However, because LMEs override all other lift blocks to be vectored, the same is true for them. In other words, for a tech with vectored lift, lift itself is not governed by the stabilization computer. I need to do an experiment to verify, but that could influence roll/pitch also.

Does that mean that we can (re)mark this thread as "[Resolved]" then? As there's no glitch/bug/code issue here.
Yes, fine. It's "resolved" but I still think there's a game design problem here. The stab computer still has limitations that I don't think were intended by the devs, even it is working as implemented. I doubt they worked through this level of physics when creating the stab computer.

But the gravity vector is always straight down towards the ground, while any small purturbation in the tech's position will offset the force from the anchor tether from being 180degrees opposed. It's all about finding stability *despite* disturbances to position and rotation of the tech, which requires negative feedback or some type.
Yep, this is pretty much exactly my thesis statement in your own words. The force-system of lift, gravity, and the anchor guarantees you will always have some force perpendicular to the anchor (or "sideways"), no matter what you do. So the key is to reduce its effects to the point that other parts of your design can control it. Anchor spacing and more gyros counteract the force (if indirectly), while moving anchors closer to the center-of-mass reduce the magnitude of the force. (Keep in mind, I'm interpreting the anchors as a force vector(s), which is physically true, but may not be self-evident.)

As for spreading the anchors out, I 100% agree with you. But once again, it's a matter of design. Your anchors can be closer together if the base stability of your tech is higher. In the picture of my airship, the back anchors are actually quite narrow, but because it's so close to the hairy edge of stability, it's sufficient. But I agree in principle, the wider the anchors, the better the stability. The increase in anchor height will also make it all stability issues worse, but hopefully the increase in gyro strength will help.

Incidentally, I noticed (by testing with a free floating AG tech) that the forces exerted by the tethers are a little weird - they actually provided zero resistance up to 40m vertically, while immediately providing resistance to any horizontal offset:

I think that they probably do add a physically logical a downwards force component with their sideways forces, though. But tricky to be exactly sure.
I didn't realize they provide horizontal force. Hmmm...
 
Last edited:
Likes: Matt