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

Saelem Black

Tinker, Tailor, Soldier, Spy
Joined
Aug 25, 2018
Messages
321
Likes
945
Points
205
#1
Why doesn't the stabilization computer work while sky anchors are down? I suppose I've never seen a reason for they don't - it seems like a common sense thing. Right now the only way to keep a skybase up with a huge number of gyros, especially since gyros are so weak on large techs. It would be dramatically easier if the stab computer could control a few adjustment thrusters.
 

Matt

PAYLOAD STUDIOS
Joined
Jul 23, 2018
Messages
431
Likes
1,282
Points
305
#2
The stabilising computer just auto controls things that provide thrust to try and keep you still. It works with hovers, adjustment boosters and propellers.

It won't work with sky anchors, because sky anchors don't provide thrust, they just stop you moving too far away.
 

Saelem Black

Tinker, Tailor, Soldier, Spy
Joined
Aug 25, 2018
Messages
321
Likes
945
Points
205
#3
The stabilising computer just auto controls things that provide thrust to try and keep you still. It works with hovers, adjustment boosters and propellers.

It won't work with sky anchors, because sky anchors don't provide thrust, they just stop you moving too far away.
Let me clarify. When you anchor (with sky anchors) it shuts down your stabilization computer altogether.

In other words, as soon as you anchor your skybase or airship, all stabilization control is lost. You still have direct control over your thrusters/lift mechanism but they no longer passively stabilize. When you're anchored, it's as if there's no stab computer attached at all (and you have to trust your gyros to do all the work).

Was this intentional? I always assumed it was. It's been this way as long as the stab computer has been around. I noticed it first when building that comcore hover train from a few months ago.
 
Last edited:
Joined
Jun 24, 2017
Messages
1,858
Likes
4,295
Points
525
Age
17
Location
Near a duck pond in Britain
Website
duckduckgo.com
#5
I've checked this with one of my techs, and it seems that the stablisation computer does activate with a sky anchor on, but it seems to like throwing the tech around for some reason. Below is the tech I used for testing, I pressed W and when I released the key, the steering hovers did try to stop the motion of the tech.
 

Attachments

Likes: ZeroGravitas

Saelem Black

Tinker, Tailor, Soldier, Spy
Joined
Aug 25, 2018
Messages
321
Likes
945
Points
205
#6
I've checked this with one of my techs, and it seems that the stablisation computer does activate with a sky anchor on, but it seems to like throwing the tech around for some reason. Below is the tech I used for testing, I pressed W and when I released the key, the steering hovers did try to stop the motion of the tech.
I'll have to test this some more when I get home. It's quite possible that what I'm observing is some compound effect. However, I have several airships (my profile pic, for one) which can hold position very precisely while *not* anchored. The second they anchor, however, they lose the ability to passively stabilize and end up rolling over and falling. I noticed when making my hover train (quite a while ago, and worth a double check) that the adjustment thrusters were no longer firing as soon as the sky anchors went down.

Keep in mind that gyros always work regardless of anchor-state, so that might also be what you're seeing.
 
Joined
Jun 24, 2017
Messages
1,858
Likes
4,295
Points
525
Age
17
Location
Near a duck pond in Britain
Website
duckduckgo.com
#7
I'll have to test this some more when I get home. It's quite possible that what I'm observing is some compound effect. However, I have several airships (my profile pic, for one) which can hold position very precisely while *not* anchored. The second they anchor, however, they lose the ability to passively stabilize and end up rolling over and falling. I noticed when making my hover train (quite a while ago, and worth a double check) that the adjustment thrusters were no longer firing as soon as the sky anchors went down.

Keep in mind that gyros always work regardless of anchor-state, so that might also be what you're seeing.
I was staring at the steering hovers and they were definitely producing the hover particle effect, so I'd say the stabilisation computer is at least doing something.

I do however remember placing a sky anchor on one of my airships when the anchors were first released, and the airship was definitely being tossed about like it was being punched, but I don't remember if it had a stabilisation computer or not. I'll do some more testing of my own sometime.
 

Matt

PAYLOAD STUDIOS
Joined
Jul 23, 2018
Messages
431
Likes
1,282
Points
305
#8
So the Stabilising Computers are meant to work while anchored with Sky Anchors.

If you suspect they aren't working as intended we can post this as a bug report and get @kae to take a look.
 
Joined
Jun 24, 2017
Messages
1,858
Likes
4,295
Points
525
Age
17
Location
Near a duck pond in Britain
Website
duckduckgo.com
#9
I did some testing. Here, I used one of my airships (Dire Wolf MKII) and did a side-by-side test, one with a computer, the other without. The one with the computer stayed stubbornly in place even when I tried knocking it around, while the one without fell over on its own within 10 seconds. A snapshot of the tech used is available below.

I have a theory on this:

From my own experience with them, the stabilisation computer seems to favour using hoverpads for stabilisation with fine movement, and only uses fans/adjustment thrusters etc to correct large movements like swaying and inertia. In the video above at 0:52, this can be seen when I knocked the airship for a second time from the back. It moved, but the fans did not react. This behaviour can explain the different results Saelem Black and I got.
As far as I know, Saelem only uses fans and steering hovers/adjustment thrusters to lift and fly their airships, so the stabilisation computer can only stop large movements and cannot prevent smaller movements from happening, causing the airship to topple eventually as the fans overcompensate and cause extreme swaying.
My airships use hoverbug and fans for movement and lift. The fans stop large movements, while the hovers fine-tune the position of the airship with smaller movements, keeping it stable and stopping the fans overcompensating.

I think this issue needs more investigation.
 

Attachments

Last edited:

Matt

PAYLOAD STUDIOS
Joined
Jul 23, 2018
Messages
431
Likes
1,282
Points
305
#10
One thing to note is the Stabilising Computer only works with hovers that have vectored thrust which, as far I can tell, is only Better Future.

Does this information help in any way?
 

Saelem Black

Tinker, Tailor, Soldier, Spy
Joined
Aug 25, 2018
Messages
321
Likes
945
Points
205
#13
I did some testing. Here, I used one of my airships (Dire Wolf MKII) and did a side-by-side test, one with a computer, the other without. The one with the computer stayed stubbornly in place even when I tried knocking it around, while the one without fell over on its own within 10 seconds. A snapshot of the tech used is available below.

I have a theory on this:

From my own experience with them, the stabilisation computer seems to favour using hoverpads for stabilisation with fine movement, and only uses fans/adjustment thrusters etc to correct large movements like swaying and inertia. In the video above at 0:52, this can be seen when I knocked the airship for a second time from the back. It moved, but the fans did not react. This behaviour can explain the different results Saelem Black and I got.
As far as I know, Saelem only uses fans and steering hovers/adjustment thrusters to lift and fly their airships, so the stabilisation computer can only stop large movements and cannot prevent smaller movements from happening, causing the airship to topple eventually as the fans overcompensate and cause extreme swaying.
My airships use hoverbug and fans for movement and lift. The fans stop large movements, while the hovers fine-tune the position of the airship with smaller movements, keeping it stable and stopping the fans overcompensating.

I think this issue needs more investigation.
Really good insight. How I've been evaluating this is with the attached tech (snap may or may not work. I'll update it in a few hours with a working one for sure, you can grab a working one on my techs thread post #16 if this one doesn't work). It was designed earlier this year when gyros were much stronger. At the time, it had no issues staying stable while anchored. Then gyros were nerfed (and nothing else). It's performance didn't change at all in air. Still completely stable. But it can no longer keep itself stable while anchored. To be fair, this uses predominantly hawkeye exfil thrusters for stability and LMEs for lift. Maybe the exfil thrusters don't play nice with the stab computer?

My logic is that the gyros only were what was keeping it upright while anchored (which was more than enough when gyros were super powerful). Otherwise, why would it be stable while unanchored, but unstable while anchored? The adjustment thrusters do their job fine as long as the anchors aren't down.
 

Attachments

Last edited:
Joined
Jun 24, 2017
Messages
1,858
Likes
4,295
Points
525
Age
17
Location
Near a duck pond in Britain
Website
duckduckgo.com
#14
Really good insight. How I've been evaluating this is with the attached tech (snap may or may not work. I'll update it in a few hours with a working one for sure, you can grab a working one on my techs thread post #16 if this one doesn't work). It was designed earlier this year when gyros were much stronger. At the time, it had no issues staying stable while anchored. Then gyros were nerfed (and nothing else). It's performance didn't change at all in air. Still completely stable. But it can no longer keep itself stable while anchored. To be fair, this uses predominantly hawkeye exfil thrusters for stability and LMEs for lift. Maybe the exfil thrusters don't play nice with the stab computer?

My logic is that the gyros only were what was keeping it upright while anchored (which was more than enough when gyros were super powerful). Otherwise, why would it be stable while unanchored, but unstable while anchored? The adjustment thrusters do their job fine as long as the anchors aren't down.
So I tested your tech, and it indeed does fall over for some reason when anchored. I tried this with one of my helicopters and the same thing happened. I believe my initial theory is wrong and that the fans and adjustment thrusters are actually undercompensating, because they only become active when the tech is tipping quickly, so they come on too late to stabilise the tech.
 
Likes: Saelem Black

ZeroGravitas

Breaker of Games
Joined
Jun 29, 2017
Messages
3,636
Likes
4,845
Points
675
Age
36
Location
UK
Website
www.youtube.com
#16
Sounds like we have a bug.
I don't think there is, after having taken a good look. But this is clearly a bug discussion, so I'll move the thread (from "suggestions") and rename it accordingly (previously "Make stabilization computer operate with sky anchors").


I've played around with a few test platforms, like this snapshot (also one with wheels and one with hovers, instead of the LMEs) and for all of them I see all of the fans, adjustment thrusters and hover plates being correctly controlled by the Stabilising Computer.

StabComp Test 1b.png

Instructions: load snapshot, anchor, set LMEs to max with lift up, fire boosters to stretch air anchor, allow the teather to pull the tech back into position, observe all thrusters, etc, pushing against the techs movement (while you press no control keys).


As Quack and I discussed in another bug thread, there is an issue with the Stabilising Computer muting the amplitude of the vectoring of hover pads when using a powered up throttled control axis (and there's no user input) - seen as a discrepency between speed of the demo tech with/without computer (and vissibly less deflection of the hover rings):
There's a third factor to cause this bug: throttled thrust control; it only happens on a control axis with HE rotors or BF LMEs. (Which will be why I've not spotted this before, because I never use the clunky throttled controls.)

This tech demonstrates this more clearly (by setting forward thrust high, then removing/adding the computer):


I don't think anything either of you have said above indicates you were seeing anything like this. Quack's demo just showed that the computer works fine while air anchored, of course.


At the time, it had no issues staying stable while anchored. Then gyros were nerfed
And, in short, that's the whole reason for the change to it falling out the sky when anchored, I believe.

In long:

(1) That big LME airship is very finely balanced for lift, tip it a few degrees sideways and the vertical component of the lift vector is no longer enough to maintain altitude.

(2) It also has so much inertia that it slips sideways too slow for the computer to fire up any compensatory thrust. This is function as intended; computer is not designed to correct for location drift, it only counteracts speed, aiming for zero (or rather, less than ~1mph 2mph).

(3) Additionally, everyone builds their sky bases, etc, with the sky anchor slung at the bottom, which is a bad plan. This means that any restorative force is applied sideways to the bottom of the tech, like pulling the rug out from under someone, it will cause it to tip away even more!

The best plan is to locate the sky anchors higher up, inside the tech, in line with the centre of mass. This way no rotational force is applied. I've made this modification (quick and dirty) to your snapshot tech, Saelem, and it floats anchored just fine now. (Incidentally, it's even worst to have the sky anchor attached high *above* the centre of mass, because it sets up negative feedback oscillations that can easily get very out of control, lol!)

ZG SaelB AnchorRe1.png
 
Last edited:

Saelem Black

Tinker, Tailor, Soldier, Spy
Joined
Aug 25, 2018
Messages
321
Likes
945
Points
205
#17
@ZeroGravitas , okay, I follow you, but that still doesn't explain to me why my ship stays stable without anchors down. If I'm reading everything you said correctly, it should have the roll problem regardless.

(2) It also has so much inertia that it slips sideways too slow for the computer to fire up any compensatory thrust. This is function as intended; computer is not designed to correct for location drift, it only counteracts speed, aiming for zero (or rather, less than ~1mph).
But the stab computer clearly counteracts drift while unanchored, so my question still stands I think. Your StabComp tech is also somewhat suspect to me because you have a HUGE amount of total gyro force on it considering the overall weight. With the 1.1 Stable gyro strength, my airship was stable, too.

The easy fix for my ship is to cram a bunch more gyros on it, but I still need to understand why being anchored changes the stability.

Also as my thread, I'm changing the title back to being a little less conclusive until I fully understand what's going on. No offense, but it's not clear at all to me that everything is working as intended.
 
Last edited:
Likes: Aardvark123

Saelem Black

Tinker, Tailor, Soldier, Spy
Joined
Aug 25, 2018
Messages
321
Likes
945
Points
205
#18
Okay, I think I figured this all out. Thanks to @ZeroGravitas for being a sounding board while I thought through this. It's not that the stab computer isn't working, it's that it 1.) doesn't do pitch/roll correction (because center of mass stays at "zero" velocity) and 2.) doesn't resist motion less than 2 mph.

This is my math proof of how this happens; more for me than anyone else, but feel free to give it a read to see if there's logical consistency.

1.) You start out with an airship tech, level and in the air at some low elevation.

2.) You anchor, but the anchors are not fully extended yet.

3.) The nature of vectored lift demands that you're either above or below hoverpoint. You can never be there exactly. Thus, to stay airborne, you throttle up (even a tiny bit above hover).

4.) At max anchor distance, the anchor exerts a force to keep your tech a fixed distance from the ground.

5.) If there is any (and I mean literally any) horizontal motion, less than utterly perfect tech weight balance, or less perfectly level terrain, there will be some tiny force vector in a horizontal direction. This is the vector which is the resultant of the force triangle between gravity (straight down from the Center of Mass) and the anchor beam (almost straight down, and from the anchor block). Moving the anchor deeper into the tech, closer to the center of mass minimizes this vector, though it doesn't eliminate it.

6.) This tiny horizontal force vector pushes you to the side with some small speed. It is too small for the stabilization computer to counter.

7.) Now this is the important one. If you're unanchored and mid-air, this tiny vector makes you slide with negligible velocity. You're still moving, but so slow it doesn't matter. However, when anchored, you're tethered to a single point, enforcing a max distance from the ground to your tech. As a result, this sideways drift eventually forces you to begin rolling/pitching.

8.) The stab computer doesn't recognize angular motion around your center of mass, so it doesn't perceive anything is wrong. The tech's total velocity is still too small for it to read, and you begin pitching more and more sharply.

9.) Gyros will stop the pitch eventually... if there's enough gyro force. If not, eventually the component of downward force produced by your lift blocks (which is the cosine of their total lift), will overwhelmed by the tech's mass.

10.) The tech falls.

Important conclusion: I don't think it will matter what the tech's weight is. It's about the gyro force being able to control the pitch/roll angle (which is indirectly related to mass, to be fair). I think this will happen to any tech without enough gyros.

So, ZeroGravitas, imo this is yet again an issue that comes down to gyro scaling.
 

ZeroGravitas

Breaker of Games
Joined
Jun 29, 2017
Messages
3,636
Likes
4,845
Points
675
Age
36
Location
UK
Website
www.youtube.com
#19
Geez @Saelem Black, you're really having a good go at trumping me on verbosity... :p;) Everyone else has probably stopped reading this thread by now; what a pair, heh. :rolleyes:

his is yet again an issue that comes down to gyro scaling
Not really. In that you can make stable sky anchored hover bases/platforms/etc with zero gyroscopes on them (I'll demonstrate below).

But yes in the shallower sense, that a reduction in gyroscope force made the difference between your specific ship design(s) being stable when anchored and then falling after the patch change (as we've affirmed several times already).

But my thinking is that you should always avoid using gyros as a bandaid for badly balanced craft. This applies almost as much to this airship tethering as to airplanes which inherently nose up/down because they have too few wings at front or back or too much weight.


Your StabComp tech is also somewhat suspect to me because you have a HUGE amount of total gyro force on it considering the overall weight.
That was purely a stabalising computer operation demonstrator tech, in no way intended to test/show airship balance.

I don't think it will matter what the tech's weight is.
Yes, you're right. I shouldn't have mentioned high mass, as that really just determines the speed and manner in which it fails, not determining the presence of instability itself. You kinda focused in on the least relevant part of one of my 3 points there - the lack of positional stabilisation (from the stab comp) being the essential bit of (2). That it can only slow the tip and fall, not prevent it.

And the most important factor, overall, being (3): the position of the anchor relative to the tech's centre of mass. I think you're getting this, from talk of angles and component forces. But here's a demonstrator for that, anyway (I feel I could do with making a proper video tutorial some time for this):


(a) With anchor block near tech's centre of mass (CoM) - stable with just 2 gyros (no computer).
(b) With anchor low down below CoM - tips over (twice). Kinda like trying to halt someone on roller-skates from rolling down a slight hill, by holding onto their their ankles. The leverage
(c) Anchor above CoM, the rotational effect on the tech tips it back towards the direction it came from, naturally stabilising position, but causing large oscillations, as the motion is undamped (very little friction).
(d) Etc, removing gyroscopes, progressively, to show (semi) stable sky-anchoring without them. Stab comp or air brakes can help dampen oscilations.

I was going to draw you a free body force diagram, which would have been worth more than all these words, but I've too tired for today.

Snapshot, to test it out first hand:
SkyAnch Stab 1.png
 
Last edited:
Likes: Aardvark123

ZeroGravitas

Breaker of Games
Joined
Jun 29, 2017
Messages
3,636
Likes
4,845
Points
675
Age
36
Location
UK
Website
www.youtube.com
#20
I'd noted the 2mph (not 1mph) limit on the stab comp's stabilisation previously, e.g. here in this suggestion thread:

Fans unable to hold tech on a slope despite being easily powerful enough. (Perhaps desirable?)


(5) Station holding - Will cause the tech to actively seek the X,Y,Z coordinate you left it at. Applying gentle forces to that end.

Takes all the strain out of station holding while bombarding a target with an airship/helicopter. And perfect for making floating sky bases.:)
And it's also worth noting that the stab comp *does* counteract rotational motion (but again, not absolute pose, of course). And you can see that by spinning one of these demo techs and then watching the thrusters or hover rings being fired up in reverse, when you let go of all the controls:

Test spin 1.png
Test spin 1b.png