Let's Make Robots!

Servo selection problem

I have been working on a robotic arm for my first solo project from scratch. Everything has gone well for the most part, I will post the robotic arm when it is finished. I have run into a problem though and need some help with what to look for in a servo as far as specs. The problem is, the first vertical joint cannot handle the weight of the rest of the arm very well when it is idle. In motion, the arm swings fine, when at rest, it has a tendency to just fall over. I am assuming I need to buy different servos for use in my arm. I am currently using mostly futaba s3003's, except the swivel base which is an s3001, which is more powerful.

What kind of servo do I need to buy or ask for that will support at least two or three more servos worth of weight at rest?


Am I going about this the wrong way... Do I need to double up my servos at the lower joints? Or maybe I am going about this the completely wrong way.

Any help would be appreciated or a link to a guide to picking out servos(I tried searching the site for such a guide but no luck).


Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

Remember, you know what your robotic arm looks like and what you want to use it for. We do not. Please post pictures and more info to help us help you.

The robotic arm made out of styrene mostly. It is not very big. Maybe 2 feet long(about 2/3 of a meter). Each portion of the arm is about 6 to 8 inches long. The material that makes up each portion of the arm weighs considerably less than the servo it supports. The base, which rotates horizontally(like a lazy susan) is the first servo futaba s3001. The first joint is just above the base, rotates verically, it is a futaba s3003 and is giving me trouble when at rest. It seems it cannot support the rest of the arm when it is not moving, however, the amount of torque it produces when it does move is fine. It has no problem lifting the rest of the arm, it just tends to fall over when it is not moving. The next joint is about six inches further and could present the same problem as the first joint once I add the weight of the claw.

Most of the weight of the claw is made up of the servos. It is intended to be a light duty arm. I don't need it to lift any thing of any significant weight or even be that cool. I'd be happy if it just looked as if it could lift something, but really couldn't ;) I am just seeing if I can do it. Proof of concept I guess. I know my description may not be enough help, but I was planning on stopping at the hobby shop on the way home, I was hoping I could get some help, because the shop is kind of far from my house. If I go home and take pics and post them, I will have to wait till tomorrow to get the parts and I wanted to work on it tonight. Thanks you.

Hmm. Since you only need some stabilization of the lowest vertical joint when it is still, have you considered adding some springs?

The servo will have to be strong enough to overcome the spring tension when it moves, but it may be just enough to keep it still when at rest.

Another thought: If you unloaded the vertical joint servo (disconnected the arm), would it still move a bit or is it rock solid? Perhaps the servo is faulty, affected by noise, or just not accurate at keeping station. Observe the loaded/unloaded behavior and decide if this is a possibility.

Yeah, i don't know how I would go about setting up the springs. But you understand the dillemma. I need some way to lock the joint in place when it comes to rest. This may be beyond my ability though :(

I am judging this by other arms I have seen...

At 2 feet long, you will need some pretty major servos for the base. I have seen small arms -- much less than 1/2 meter fully extended and they are using 2 heavy-duty servos at the base and running a full 6v from a separate supply. For an arm 2' long, unless very very light will probably need 1/4 scale servos.

Here is how to calculate a servo's strength: Say a servo is rated a 1kg/cm --this means if the servo had a little arm on it and that arm was 1 cm long, the servo could lift 1kg. If that arm were 2cm long, it can lift 500g. At 4cm we are down to 250g. Simply weigh the arm including everything except the very base servo, measure it's length and from there you can calculate easily what you are going to need for the base. Don't forget to add the gripper's load as well to the equation. I think you will quickly find you need some very beefy servos.

I always wondered how those weird kg cm units worked, thanks.

the thing is, the servo has no trouble with the weight when in motion. Only when it is not moving, if the weight isn't balanced, if the arm is leaning slightly, it will not support itself. The weight will actually move the servo when at rest. Like I said, it does not struggle at all when in motion. The arm is very very light weight, made out of very thin poly styrene. The bulk of the weight is the servos themselves.

That is, odd.

It could be a power issue, How is this powered?

A video would be helpful.

What controller are you driving these servos with? How frequently are you issuing the position commands?

Thanks for your help everyone. It was a power issue. I guess I made a mistake during the troubleshooting session last night when I was messing with it. When the servo has power to it, it supports the weight even when still. Sorry for the confusion. I could have sworn it was falling regardless of it having power or not when I was messing around with it. As it turns out, whenever i wish to turn the arm off, I am going to have to return it to a position of balance. Thanks for the help, will post some pics once I get a little more done.