I have been modelling the B9 robot from Lost In Space over the past week or so.

I was wondering if anyone would have any ideas on how to best rig his arms? If the arms were strictly mechanical in nature, than that wouldn't be too much of a problem. But since his arms were essentially flexible ducting with a human arm inside, it was very flexible in nature. It could be fully extended or retracted, and when extended could move around with total freedom.

Any ideas or suggestions would be greatly appreciated.

Below are some pictures of his arms in 'action'.

Thanks...
Al

Extended Arms
Click to view attachment

Retracted Arms
Click to view attachment

Arms Extended Straight
Click to view attachment

Arms Extended To The Side
Click to view attachment

You've Got Me... Drunken Robot??? Arms Completely Limp
Click to view attachment

For starters, I would try to rig the arms just like normal humanoid arms, except don't make the claws (the bones assigned to the claws) children of the arm bones. Put them at the top level of the bone tree. Then put a "translate to" constraint on them to keep them attached to the arms. It helps to add a small "placer" bone to the end of the forearm bone (make sure it is attached to the forearm bone). It will be easier to target the "placer" bone in the "Translate To" constraint .
Then make a pose to retract the arms. In the pose, scale the upper arm bone along the Z axis.

If the claw bones are children of the arm bones, then they will scale too when you scale the upper arm bones, which is probably not what you want.

I would be interested to see other ideas on how to rig the arms. I'm sure there are other ways.

For TWO I did a concertina with pleats much like the robot's arms

Click to view attachment

Click to view attachment

Every rib of the pleats has a bone and then the bones were single percentage path constrained along a spline that ran thru the middle of the model.

Thanks guys for your suggestions.

I had already tried scaling the bone, but found out exactly what you said here Holmes. The wrist and claw bones scaled as well.

I've been playing with the example you provided Rob, and it works really great! I've spent this evening getting it to work on one arm, but have been having trouble getting two arms to work. I tried the mirror bones, but it wasn't mirroring the relationships with it (not sure if it is supposed to or not). I am now in the process of rebuilding the relationships for the second arm. Once I get something working, I will post an update.

Thanks once again guys. Your help has been invaluable.

Al

This post is probably too late, but if Rob's model has very many relationships, you can save a little time by:
Make a copy of Robert's model.
Rename the bones in the copy (add a 2 after each bone name) - the relationships should update with the new bone names.
Save
Drag the copy into your robot model.
Now you have working relationships for both arms.

I think at one time I experimented with a middle controller on the concertina to more explicitly control the curve, but I guess I dropped that notion by the time I got to that version.

As it is, you have to control the curve by the orientation of the endpoints, which may be inadequate for the robot since the body end will be fixed.

Robert

I took a look at the rigging and relationships in the model. Very impressive. How did you come up with that? Can you make a tutorial of the theory behind it?

For an easy aproach:

How about using a part rig from TSM 2, like a tail rig for each arm?

I had tried that. Once I got the right arm working (with all the bones and groups labeled 'Right whatever'), I saved the model as both right and left arm models. Then I loaded the left arm model, and renamed everything 'Left whatever'. I repositioned the left arm and bones to the appropriate place, and saved that model. Individually, both arms worked perfectly. When I tried to merge the two arms into one model or in a chor, the arms would go all screwy and stop working. In the chor, it looked like control points from the left arm were merging and being influenced with the right arm and vice versa. I tried redoing this 4 or 5 times different ways, and it would always end up screwy. I think it came down to the fact that I wasn't making a copy of the splines and control points, but using the same ones from the right arm, but in a left arm model. I ended up (late last night), making an actual copy of the splines and CPs, flipping it along the X axis, then mirroring the bones. I am now in the process of rebuilding the relationships. This does appear to be working, but it is slow going (I am not the faster rigger... )



Not a problem Rob. I really appreciate the work that has gone into your concertina and thank you for providing it. It really is a brilliant piece of work. It is taking a bit, but most of how it works is slowly sinking into the ole' noggin. I have done some initial testing with moving one end of the arm, and it works great. I think I may try adding a NULL so the end bone is associated with it. The NULL may make it easier to grab ahold of and move the arm around.



I don't have the TSM 2. Would it allow for extending and contracting as well as moving the arm around?

Thanks again...
Al

Plan A, B and C didn't work.




I'd have to study it to remember all the details, but really it's about constraining all the bones for the ribs to the spline path that runs thru the model center. Each one is given a progressively larger "ease %" setting so it holds its particular station along the length of the spline.



I don't think a TSM tail would do the extending and contracting.

TSM2 is free and available here. I have a tail rig that would more than do the job, but it would be overkill (I'm hoping to post it this week, but I'm not sure when it will be ready for showing yet). A tweaked version of Robert's concertina rig would probably be the way to go.

Got the arms done, and posted a test animation in the WIP section.

Robot WIP Test Animation

Thanks again everyone for your great help...

Al