Author Topic: My first robot project question  (Read 626 times)

0 Members and 1 Guest are viewing this topic.

Offline darthmapleTopic starter

  • Beginner
  • *
  • Posts: 2
  • Helpful? 0
My first robot project question
« on: July 27, 2013, 08:54:11 PM »
I'm in the first phase of designing my project. so far i know that it will be bipedal and about 5-6ft tall. in other words its gonna be similar to a human. im going to make the frame first out of tubed steel for the sake of simplicity. I need to know what exactly i will need after i get the frame built.

Offline MrWizard

  • Full Member
  • ***
  • Posts: 116
  • Helpful? 0
  • My cylon friend told me a killing joke......
Re: My first robot project question
« Reply #1 on: July 28, 2013, 01:56:53 AM »
First of all making a scope, what you want to do with it. Calculate the forces on the skeleton/body for moving. Which implies what material you are going to use. Then try to figure out how many joints. So you can select the motor-type and how many motors. Then you look at the microcontroller, how many motors do you want to control and ofcourse which sensors, the battery-type. Select the stores and do not forget the junkyard. Buy, build or scrap the parts. Start mounting and start programming. Make lots of fotos and movies to keep us happy advising.

Offline darthmapleTopic starter

  • Beginner
  • *
  • Posts: 2
  • Helpful? 0
Re: My first robot project question
« Reply #2 on: July 28, 2013, 09:32:43 AM »
Thanks for the help man. which type of motor would be best for elbow joints and which is best for shoulder joint
« Last Edit: July 28, 2013, 09:39:59 AM by darthmaple »

Offline jwatte

  • Supreme Robot
  • *****
  • Posts: 1,343
  • Helpful? 81
Re: My first robot project question
« Reply #3 on: July 28, 2013, 10:38:48 AM »
I think steel tubing will be too heavy. The main problem in all bipeds is to get strong and accurate enough servos (motors) for the joints, and the strength needed is directly proportional to the weight. I would *highly* recommend aluminum tubing, with thin walls, or U-profile frames.

Second, you can calculate what kind of motors you need through measuring the lever length and the weight. The motor "working torque" needs to be at or above this product to avoid overheating. If you can only find "stall torque" listed, then "working torque" is often at 20-25% of stall torque.

Make sure to include the weight of the servos themselves, wiring, any fastening hardware, etc. So, for example, for a 6' humanoid, the lower arm + hand might be 20" long. Assume hand effectors with a weight of 2 pounds, and a frame + wiring weight of another 2 pounds. Assume center-of-gravity of that assembly is 15" out from the joint (as the hand pulls it further out from the center of the assembly.) This means you need working torque of 1.25'*4lbs == 5 ft-lbs, which turns into stall torque of 25 ft-lbs, or about 4,800 oz-in, or about 34 Nm.

Now, you can go to Kollmorgen or Maxon or Faulhaber or Bosch or GE or Sumitomo or whatever industrial motion control company you prefer and see if they have a servo motor strong enough for your need. You will typically want a gearhead DC motor, for ease of control and power -- AC motors are generally bigger, bulkier, and not easy to drive from battery packs or robot-mounted controllers.

Note that you don't just need a "motor," you actually need a motor, a gearbox, an encoder (to know where the arm is,) a motor controller (to control the current to the motor) and a motion controller (microcontroller or computer that reads the encoder and tells the motor controller how to do work.) This together make up a closed look servo system.

For smaller robots, companies like Robotis (Dynamixel) and Dongbu (HerkuleX) make servos in a package that contain all of the needed sensors and control electronics together with the motor. For about $300 you get a servo with about 7 Nm of stall torque, which is a very fair price compared to what it would cost from an industrial motion control company. In general, the cost of motion control goes up as the square of the load, so going from 7 Nm to 35 Nm would give you an increase of 5x in load, and thus 25x in price. Unfortunately, there is no "hobby" level provider for servos that big, so you'd probably get another factor 2x for going industrial. So, your one-degree-of-freedom elbow joint would cost you about $15,000. As you move up towards the body, and down towards the knees and ankles, the needed torques go up even more, as they have more load to move around.

For reference, 1' humanoids like the Hovis and the Bioloid Premium are about $1,000 in kit form, with all the engineering/calculation already done for you. A 2' humanoid like the Darwin-OP is about $10k, again in kit (or assembled) form. I suggest you start with one of those to get a feel for what building and working with robots are really like.

Still want to build a 6' humanoid as your first project? :-)

Please post pictures!
« Last Edit: July 28, 2013, 10:43:02 AM by jwatte »

 


Get Your Ad Here

data_list