Inflating causes player to spin

[toxic relic]

Yo! I have a project from school where we have to make a game using one input. So I was thinking about replicating inflating from little slime blows up. I analysed the video and some more videos about soft body and I got my player to inflate! But I got some problems:
1: When I inflate, sometimes the player just spins.
2: Inflating while moving the character also spins the player.
3: I'm not sure about whether using rb.AddForce(f, ForceMode2D.Impulse) or rb.AddForce(f, ForceMode2D.Force).
Should I calculate Vector2 dir = (rb.position - center).normalized or do something with the radius?

At the start of the game I spawn circles with rigid body 2d, circle colider 2d and springs. Then I connect the springs.
When the player presses the space bar, bJumpButtonPressed is set to true and then in fixed update I calculate the direction vector:
I get the position of the rigid body of the circle - the centre point of the player (I have uploaded a picture of this vector * 10 in the red line).
and as long as you hold the space bar, you stay in the inflating state.
When you release, the circles should stop inflating and return to where they were before. (which they do)

The problem? Sometimes the player just spins itself, causing the inflating to move the player left or right (which I don't want). And that's without mentioning the fact that I'll need to move the player right which makes the spinning even more problematic

Scripts:
Spawning the circles and connecting the spring joints: https://paste.mod.gg/ohhzsonauhds/0
Inflating the player and moving it: https://paste.mod.gg/gcqnreqyvtie/0

[heady copper]

A video would be super helpful. But spinning while moving here is likely related to friction

a ball-like object such as this will tend to roll

[toxic relic]

@heady copper Here is a video

[heady copper]

Could you share your code?

[toxic relic]

I know that there are ruls here to sharing code right?

like with !code

!code

[peak laurel]

Posting code

📃 Large Code Blocks

Use links to services like:
https://paste.mod.gg/, https://hastebin.skyra.pw/, https://paste.ofcode.org/, https://paste.myst.rs/

📃 Inline Code

Surround code with three backquotes. Not quotation marks.
To format as C#, add cs to the first line:
```cs
// Your code here
```
Add a comment with a line number if there is an error message.

[heady copper]

Yeah read this^

and share your code like that

[toxic relic]

https://paste.mod.gg/ohhzsonauhds/0

[toxic relic]

Spawning the circles

https://paste.mod.gg/gcqnreqyvtie/0

[toxic relic]

Movement

I have coding experience with ue5 (both C++ and blueprints)

So if my practices are not perfect then please tell me so I can write better code

[toxic relic]

@heady copper Do you know what's the problem?

[heady copper]

I think at some point just due to floating point inaccuracy you're pickingup a slight tilt to one direction or another, and then the expansion force is amplifying that. And then your code uses the average position of the circles to calculate a lot of stuff which gets really thrown off especially when part of the whole thing is compressed

It's hard to say - but I think you have a lot of small floating point error that gets amplified and thrown back through the feedback loop

Also this:
rb.AddForce(delta * (50f * Time.fixedDeltaTime), ForceMode2D.Force);
is wrong in a couple ways

1. Time.fixedDeltaTime should never be multiplied into the force. It's automatically done so when you're using ForceMode2D.Force
2. 50f * Time.fixedDeltaTime is usually just 1, basically cancelling it out anyway. Because fixedDeltaTime by default is 1/50 (0.02)

so I recommend just getting rid of the whole (50f * Time.fixedDeltaTime) term here

[toxic relic]

@heady copper I see but that's not the inflating part the inflating part is in the first section of the function ApplyInflateAndFrameForce()

[heady copper]

when it starts rolling more quickly right here what are you pressing if anything

[toxic relic]

not pressing

[heady copper]

I really just think this is a matter of like:

• It gets some slight angle from floating point error
• then you're applying forces in or out of the center while it's moving slightly which turns into torque

and it's just a feedback loop that gets stronger

As for how to fix it I'm not terribly sure other than maybe intentionally applying your own stabilizing torque

[toxic relic]

@heady copper Do you think making the force as an int will fix this?

[heady copper]

That doesn't make any sense

[kind tangle]

Quick fix: manually set the horizontal velocity to 0 until the player starts moving. After that it's not a problem anymore because there's already sideways movement and any small aberrations disappear. When the rigidbody goes to sleep (=stops moving) start limiting the horizontal velocity to 0 again

or freeze the x axis

[heady copper]

the problem is it's a bunch of small Rigidbodies

and part of the expansion and contraction involves most of them moving sideways to some degree

[kind tangle]

hmm yeah

so maybe setting the parent x position

[toxic relic]

freezing the X axis won't do the tick becase it could mess up the inflation

Hold up

@heady copper What if we look at it from the perspective of a radius?

This is when radius is 1.05

And this is when radius is set to 5

@heady copper So armed with this knowledge, What if we can just set the radius

[toxic relic]

@kind tangle What do think?

[kind tangle]

Frankly I don't see what you're getting at or how it would help