Joint Physics Not Robust Enough to Base a Game On

Physics
#1

The central mechanic of a new game we’re developing is based on ropes. We decided to use Corona’s Box2D physics to achieve this. However, results thus far have been disappointing.

Does anyone have any tips or code examples for building a robust rope The benchmark is to replicate the ‘Cut the Rope’ aesthetic:

https://www.youtube.com/watch?v=8xPUdFaraoQ

Key markers are:

  • No separations between links.
  • No jitters or spasmodic behaviour.

Despite numerous experiments with various joint configurations and settings, it appears a very low threshold before chained joints become spasmodic. This behaviour even occurs even in Corona’s own ‘Chains’ sample code (see attached image).

All input more than welcome.

Thanks in Advance,

Ash

#2

Hi Ash,

Which type of joint are you experimenting with? I assume “pivot” joints? Can I see some of your code on how you’re assembling just one of these joints (i.e. the point where you’re joining two “links” in the rope?).

On that note, the chain should not be falling apart are completely going haywire. I’ve not observed that behavior in the “Chains” sample either, although that sample is rather old and might need some improvements. Are you applying some excessive forces in the simulation, or turning the density down to virtually zero on the objects? 

Thanks,

Brent

#3

Hi Ash,

Which type of joint are you experimenting with? I assume “pivot” joints? Can I see some of your code on how you’re assembling just one of these joints (i.e. the point where you’re joining two “links” in the rope?).

On that note, the chain should not be falling apart are completely going haywire. I’ve not observed that behavior in the “Chains” sample either, although that sample is rather old and might need some improvements. Are you applying some excessive forces in the simulation, or turning the density down to virtually zero on the objects? 

Thanks,

Brent