pic: Off Season CAD Practice

[Chris is me]

There are legitimate arguments against a roller claw in 2011, particularly for long-arm robots. A proper pinch claw can easily weigh just 2ish pounds, when a roller claw needs at minimum 1 gearbox + motor (usually 2), four shafts / pulleys, etc. Exaggerating the effect of this weight by putting it on a long arm and there's a lot more torque that the lifting motor needs to deal with. Particularly for human load robots (as this one appears to be since it doesn't seem to be able to extend past the starting config), the advantages of a roller claw aren't definitive. It's still likely the "better" choice, but many reasonable robots, great robots that made Einstein (217, 1503) were pinch claw.

None of them however were motor driven pinch claws, which are generally really bad. Motors alone can't apply force to a tube without stalling, and even when they stall the force is usually minimal. Pneumatics are really the only good way to go for a pinch claw. I would also significantly change the gripper geometry to incorporate a traction material and a more rigid structure. Your tube is already compliant - no need to make your gripper compliant as well other than to open and close it.

As for your arm design - as others have said 1x1 is a bit flimsy for this application. I would at least go with a 2x2 tube if you're set on a single piece of material. The more rigid the arm the better. Also, the live axle is definitely a mistake. You have a perfectly good sprocket with bolt holes right next to the arm, and instead of bolting the sprocket to the arm and transferring torque through multiple bolts on a 1.875" bolt circle, you're concentrating the entire lifting force of the arm on a 1/2" diameter hex shaft. Torsional failure is a VERY real possibility here and there's just no benefit to the way you're doing it now. Easy fix though - bolt your Nano to the other tower using the other face of the gearbox, leaving your output shaft in the middle instead of off to the side of the tower. You also get the advantage of being able to relieve the cantilever of the gearbox here.

I would also run the CIM through at least one other reduction, possibly two, before going to the arm. 12:1 gearing even with a big sprocket reduction is ambitiously low. I would consider something more along the lines of a final reduction of 300 or more to 1. You could do this by feeding the CIM through a 10:1 VP before going through the Nano for example.

Finally, it looks like your sprocket is extending out past the bumper perimeter of the robot. Be careful about this. You can't flush mount your tower supports to the back of your frame for this reason.

I wouldn't worry too much about the angle of the tube - people in this thread are seriously overstating the difficulty of scoring a tube at anything but the perfect angle. Teams like 1503 had no mechanism to reorient the tube and did fine, provided you design the robot so that the angle you grab from the feeder is a good angle to score on the top row you're fine.