Prismatic joints locked and weird behaviour on V 4.2.0

Report crashes, strange behaviour, or apparent bugs
Post Reply
raccarva
Posts: 4
Joined: 06 Aug 2021, 13:51

Prismatic joints locked and weird behaviour on V 4.2.0

Post by raccarva »

I had made a motorcycle front suspension simulation on an older version of CoppeliaSim (when it was called V-REP) and I tried again on version 4.2.1 but I could not obtain satisfactory simulation results.
I´ve got the following problems:
- The two prismatic joints for the suspension oleo-pneumatics seems to be locked, despite I had enabled the motor with spring-damper mode for them;
- The simulation front wheel axle is linked to the mentioned oleo-pneumatics by two force sensors with dummies but this linkage is breaking during simulation;
- The whole simulation experience seems to be weird, with lots of unexpected behaviors (where did the object get so much energy from to jump like crazy all over the scene?)

I am using the Newton engine since the Bullet and ODE are giving worse results.

If anyone can help me, the model is available for you to play with it at https://drive.google.com/drive/folders/ ... sp=sharing.

Thanks in advance.

coppelia
Site Admin
Posts: 10312
Joined: 14 Dec 2012, 00:25

Re: Prismatic joints locked and weird behaviour on V 4.2.0

Post by coppelia »

Hello,

can you tell us what version of CoppeliaSim / V-REP you where using prior to V4.2.0? i.e. in what version was the scene running fine?

But looking at how to created the model of the suspension: you are making things overly complicated. The two suspension subes can and should actually be modeled as a single suspention (since they always operate in parallel and are rigidly linked anyway). This will greatly simplify things since you won't have to use loop-closures anymore.

Have a look how the demo model Models/vehicles/motorbike.ttm is built to get an idea.

Cheers

raccarva
Posts: 4
Joined: 06 Aug 2021, 13:51

Re: Prismatic joints locked and weird behaviour on V 4.2.0

Post by raccarva »

Thanks for the response and advice!

I checked the date of my model's first version and it is 08/Sep/2018. By looking at the version history, that points to V 3.5.0, which matches my faint memories of this aspect.

I agree that having the cylinders as two independent elements seems to be redundant but that is intentional because this model is derived from a 3 wheels configuration model in which the cylinders do not work in parallel but are part of a linked elements closure. When I migrated the original model from V 3.5.0 to 4.2.0, the "locked joints" problem appeared and I then created this 2 wheels configuration in order to submit the problem to the forum, in the hope that someone could make it work as it is and help me with the original problem.

I am avoiding making publicly available the 3 wheels model until I have an intended paper published about it.

Well, now my answer also seems to be overly complex...

coppelia
Site Admin
Posts: 10312
Joined: 14 Dec 2012, 00:25

Re: Prismatic joints locked and weird behaviour on V 4.2.0

Post by coppelia »

Can you post a scene that works on V3.5.0, but not on V4.2.0? in the original format (V3.5.0), otherwise we might not be able to load the newer file and/or other problems might occurs because of new features not supported in the old version.

Cheers

raccarva
Posts: 4
Joined: 06 Aug 2021, 13:51

Re: Prismatic joints locked and weird behaviour on V 4.2.0

Post by raccarva »

Unfortunately, I can not provide any scene for the V 3.5.0 because I updated to 4.2.0.

This was the sequence of facts:
- My old 3 wheels model was running fine on 3.5.0 in 2018 (I was using Vortex on that occasion) but I stopped working on it due to other commitments;
- I went back to it now and I discovered that Vortex refuses to run due to license issues, which I do not understand since it was running before and V-REP was the same 3.5.0 from 2018;
- I tried with other engines but it could not run due to the presence of complex non-convex bodies (was it running before but only due to Vortex?);
- I migrated to the newer version to see if the model would run again but to no avail;
- I decided to redesign the model, making the non-convex bodies non-respondable and adding equivalent pure bodies. This time it runs but with the "locked joints" issue and some weird movements.

Anyway, shouldn´t the 2 wheel model be runnable on 4.2.0 as it is now? Or, in other words, is the "locked joints" issue a symptom of any bug in the simulator or in the Newton engine?

coppelia
Site Admin
Posts: 10312
Joined: 14 Dec 2012, 00:25

Re: Prismatic joints locked and weird behaviour on V 4.2.0

Post by coppelia »

It appears to me that your scene was running fine only with the Vortex engine. Did you try to update the student license with CM Labs? Vortex is by far the engine that gives most of the time the best results (in a broad sense). Having a simulation run fine with Vortex is no guarantee that it will run fine with another engine too. That's generally true for any engine.

Cheers

raccarva
Posts: 4
Joined: 06 Aug 2021, 13:51

Re: Prismatic joints locked and weird behaviour on V 4.2.0

Post by raccarva »

Ok, I will check with them if I am still eligible for a license.

Thanks for the support.

Post Reply