Manipulator designs

[Max Lobovsky]

What are some general designs for joints, hinges, etc on a FIRST robot? There are numerous white papers on drive train designs, but very little information on designs for the rest of the robot. I know the game changes every year, but almost every year has a couple of things in common, so this seems like a pretty important thing to be so severly omitted.

For example, I would like to know how teams have created "shoulder" joint for a large arm that under goes large torque in the axis of movement and in other axis. Or, a "grabber" joint where two parts of the manipulator are brought closer together.

Are there any good threads, websites, or even pictures for information like this?

[sanddrag]

Originally Posted by Max Lobovsky

Are there any good threads, websites, or even pictures for information like this?

The galleries on this site and on www.firstrobotics.net are really pretty good. The pictures are mostly of whole robots but they give you a really good idea of how the arms work.

[Joe Ross]

I assume you've already looked at this: http://www.chiefdelphi.com/forums/pa...gle&paperid=25

[Max Lobovsky]

Originally Posted by Joe Ross

I assume you've already looked at this: http://www.chiefdelphi.com/forums/papers.php?s=&action=single&paperid=25

I have seen that, and it is useful, but I'm looking for more specific information about designs. Read on...

I have looked at hundreds of high resolution pictures on CD and team sites (I can't tell anything from the ones on firstrobotics.net) and I still can't answer some basic questions. For example, the most highly loaded joints should have very large diameter axles. What diameter axles do teams use and how are the axles attached to the structure. Bushings and bearings get very heavy at these sizes, what do teams do about that? Do both parts of the joint hinge freely on the axle with the drive mechanism attached seperately, or is the load bearing axle keyed, welded, bolted through, etc to transfer torque?

What I'm hoping is for some teams with particularly good manipulators to post some information on their designs.

[phrontist]

I've been looking for the exact same information Max!

Have any teams done a real articulated arm? Any whitepapers on this?

[Ken Patton]

Some comments/advice:

If you can figure out a way to use springs (I suggest surgical tubing - its lightweight, durable, cheap, easy to install, fits a lot of designs) to counterbalance the mechanism, then the motor has a lot less work to do and you can gear it for the speed/accuracy tradeoff that you need. Think about springing the arm so that its weight is compensated for by the spring.

Friction is your enemy. Make sure you try to eliminate any sliding contacts and that you use bearings where possible or properly lubed bushings if you must.

When the robot tips over or runs into something, sometimes an arm might take extreme loads. Think about what is going to bend/break in that sort of situation, and make some spares. Lots of times the shaft on which the arm pivots is going to be the weak link.

There's nothing wrong with some deflections of the arm as it is being operated, but make sure the deflections don't screw up gear alignment or sprocket alignment.

We have used both gears and chain+sprockets as arm drive mechanisms, and both have worked satisfactorily. When selecting the motor and ratio, make sure to consider the counterbalancing.

Think about the need to hold position under situations where power is on and power is off. Do you need the arm to hold position or is it acceptable for the arm to sag? Some motors may be able to hold position by themselves, some may not. Some motors (van door and Globe come to mind) may seem "tight" at first but after use they are back-driveable and will probably not hold position under load.

Think about the need for very precise positioning of the arm - do you need to add a potentiometer or encoder to provide feedback to the controller? Can you write a control loop to take this feedback signal and convert it into a motor command? (search for threads on arm positioning and I think you'll find some good info on people's preferred methods)

Hopefully some of the experienced designers/builders of multi-segment arms might have some comments for us mere mortals....

hth,
Ken

[sanddrag]

The articulated part of our arm was made from either .090 or .125 formed aluminum sheet metal. We had a half inch hole drilled through each side. Then, we bolted titanium (steel would have been fine) flanges with a 1/2" hole and 1/8" keyway over the 1/2" holes in the sheet metal arm. The stationary part of the arm was .125 formed aluminum sheet metal in a U-type shape. This had two 1.125 holes with 1/2" ball bearings epoxied in. The moving part of the arm was placed inside the U shape of the larger stationary part of the arm with a 1/2" steel keyway shaft. The shaft rotated with the articulated part of the arm inside the bearings in the stationary part of the arm. Essentially, when hanging, the whole load of our robot was placed on two 1/2" ball bearings, epoxied in .125 sheet aluminum. It worked great.

[Madison]

There are as many different mechanism designs as there are FIRST robots, and while some certainly work better than others, they're all largely based on similar concepts and principles.

Rather than know explicitly how an arm or elevator mechanism built by team XXXX works, I think it's far more valuable to gain understanding of common engineering mechanisms and find ways creating a solution for your design problem based upon those ideas.

Ken offers excellent advice to observe while designing such mechanisms, but I think it's important that your actual designs are based on knowledge of the best ways of accomplishing a task rather than facsimiles of things you've seen on some other team's robot.

Edit: It occurs to me that you may want to dig up teams' Autodesk Inventor Award entries and examine those for a great look at how they accomplished different tasks in their robot design.

[Max Lobovsky]

I have a pretty good grasp of the basic principles involved. What I'm looking for is a bunch of solutions to this problem that other people have used. I may be knowledgable to tell a good design from a bad one, but that doesn't mean I could come up with one.

Where can I find other team's Inventor Award entries? I have seen maybe 1 or 2 team sites that have them.

[Adam Y.]

I have a pretty good grasp of the basic principles involved. What I'm looking for is a bunch of solutions to this problem that other people have used. I may be knowledgable to tell a good design from a bad one, but that doesn't mean I could come up with one.

I would recommend Building Robot Mechanisms. It's nothing but solutions. Almost everything related to robotics is in that book. It includes (usually) easy to understand picures of how the devices works. A whole chapter is devoted to arms designs and the differnces between the three designs. Also Mechanisms and Mechanical Devices Sourcebook I heard is an excellent book but I am not sure if there would be any information on the specific field you are looking for. I just came back from taking a look at the copy at the library and huge is a whee bit of an understatement. The edition that I had wasn't the latest one which contained a chapter on robotics but still the amount of information is vast.

[JamesCH95]

Last year we tried to do an arm design with Potentiometers... it can be very dangerous. If the pot decides to come loose (as will happen in competition) and "wraps around" the arm thrashed violently and almost caused severe injury on more than one occasion. I don't know if that was an inherit hardware or software flaw, but I would not recommend it. As for the load bearing aspect, we used 1/2" hardened Steel dead-axles where the arm sprockets had the bearings pressed into them. That part of the design worked well and was incredibly rugged and compact. Good luck with you design experiments!