I like the idea of doing a biped because I have been looking a lot at the prosthetic legs but I think that would make the control problem a lot harder trying to balance as well as maintain a periodic gait. The primary reason I thought of using a hexapod was because of the inherent stability of having six legs. I was going to start with two alternating tripods and work from there.
I just put together the rabbit 2000 micro with the SSC-32 and I hooked up 6 servos. I am trying to learn how to sync the servos and see how fast I can get them to move. The only problem is, the machine shop at my school hasn’t come back with my chassis yet.
I estimated that with the servos and controller and the body the whole bot will weigh 3lbs.
Also does anyone know where to find some basic design aspects for those j-shaped carbon leaf springs? Most of the places I find anything are companies that are very proprietary and wont tell me anything about how it was built. I want to design smaller versions of the prosthetic spring legs for my robot, out of spring metal, but I don’t know how thick they should be or what shape they should be.
I’ll post a picture of my bot I built on solidworks as soon as I shrink it down.
Online patent searches and google may produce useful info on the carbon fiber sprinter’s leg. Being very specific about what you are wanting to do usually produces the most useful forum responses.
I like to withdraw my previous comment J-spring on biped would be a very very bad idea.
J-spring is a very unique spring. The spring stiffness, deformation and response are unique at every different load and every angle.
You may want to design your own J-spring to suit your robot. For your case, try any J-shape first. Most important is that you verify that the servo has enough torque and power to generate dynamic gait. If otherwise, you need to scrap the idea of servo in your robot.
The first attempt you may want to try is in-situ jumping. Try jumping using tripod gait, alternating between the tripod. At first it will bounce badly, having big vertical travel and impact. Once you got the right dynamic model, the body should have minimal vertical travel, while the tripod hopping and changing. Then you can try some tripod walking gait.
If the servo has insufficient torque to do even the hopping and jumping, you need to upgrade to high torque servo, or reduce robot weight.
Your leg member length is also very important. Make sure you have contigency power and torque available for your model. Able to hop does not guarantee you will have sufficient power to generate dynamic gait. Shorted the members if necessary.
Just a quick comment on the earlier statement regarding the Basic Atom…
You said that it would be really pushing the envelope of the Atom to do this.
Is there any fundamental problem with using the Atom just for locomotion control, and using an additional microcontroller for the advanced decision-making? Ie. use the extra processor to tell the Atom which way to go and let the Atom work at running the legs.
Hopefully, this lets each processor do what it does best. Nature uses this sort of problem-solving, why not us?
I believe this is a hexapod thread, so using PowerPod to create the serial control mode, it would be very easy for the second Atom to do higher level stuff while the first Atom deals with ALL of the walking stuff. 8)
J-spring on biped would be a very very bad idea. 