Need some ideas for knee joint actuators

JohnUlett

So I have been working on this design for about a week now. I am trying to design a hexapod that has dynamic motion control and can run. Right now the only goal is to make a working model that has linear springs in the legs and an off balance rotating disc to provide rotational energy to the spring system. I am trying to design it so that when the robot steps down and through the stride phase, the kinetic energy from the landing is stored in potential energy in the spring and then helps to push the robot back up through the flight phase before the next legs touch down. In designing this, I have discovered that using a single hip joint actuator is not enough because the legs would hit the ground while reseting during the opposite stride phase. I need to design a knee join to bend back while the legs are in the recovery phase.

What kind of actuators can I use?

requirements-
knee joint must stay straight during the stride phase
bend when the leg leaves the ground
straighten again just before it sets down again.

see picture!

I would like to keep this small and lightweight and energy efficient. A few problems I see right now is that I need the actuator to act quickly. and the system needs to be strong enough to support the leg through the stride.

JohnUlett

I have thought about having the knee joint with elastic on both sides and then having a servo on top of the robot with a cable that runs down through the leg and out the back of the middle of the "hamstring" attaching to the "calf" and then when the servo head twists it would retract the cable causing the joint to bend, then when the servo switches back the elastic in the joint would extend the joint to its normal position. I don't think the servo would be able to put that much torque on the joint and if it did then the elastic would have to be to weak that it couldn't support the weight of the chassis.

JohnUlett

could I use some kind of electromagnetic control system? Anyone. And is anyone interested in working on this project with me? I am mostly doing this for fun but I have a lot more drawn up and some ideas written down about the dynamic motion control system. Shoot me a message or a reply if you want to know more or have any ideas. Or let me know if you have no idea what i'm talking about and maybe I could try to explain the whole project idea.

iamdenteddisk

I'm gonna say to look inside an old diskdrive or harddrive "for not the read head or recording mechinizm but the mechinizm moveing it",· "It will look like a pendalum from a clock only the part of the weight on the pendalum is a coil of wire wich connects to a·circuit wich turn the coil on and off in efect acts as a "electro-magenet" wich will exibit polarity north/south this device is placed in proximity to another ·magnetic field beit natural or genorated and will attract or repell with respect.

meaning· the parts move because the polarity alternates or changes, this is known as a liner particle acceloration.· so If the idea is can you use it"magnitizm" to redistribute weight,the answer is yes·.
ofcorse this isnt the device you need,but with the right info you can build one.

your closer with a tube and a coil pushing and pulling magnetized lugnut's notted on a string it works well·but warned ,when you induce a field it will emit or radiate current too -"the lugnut will shock you". think of it as the secondary in a transformer .

as much as·I would like to say I like the idea I want to also steer you right, the big reasons why these arent used· in many robotics projects is because of the fields effect on other circuits·in proximation,·like poping&kracking in radio or noise in digital circuits isolation is important,let alone its effect on small motors,compass,accelorometer's,relay's misbehave and radiation,·it dose work but isolation is the key.


··

JohnUlett

I realized when I was in class that the size I am thinking about for the first prototype of this bot would not allow for the strength of magnetic field necessary as it would interfere with the other circuitry. The same concept of electromagnetic interference would also prevent me from using a larger solenoid type device for the weight distribution system and I should stick with the rotating off-balanced disc. The Rotating disk idea with a centrifugal clutch is a great idea for a larger scale bot but I don't think I have the tools or the ability to design a small enough clutch.

tedbeau

johnultett said...
I have thought about having the knee joint with elastic on both sides and then having a servo on top of the robot with a cable that runs down through the leg and out the back of the middle of the "hamstring" attaching to the "calf" and then when the servo head twists it would retract the cable causing the joint to bend, then when the servo switches back the elastic in the joint would extend the joint to its normal position. I don't think the servo would be able to put that much torque on the joint and if it did then the elastic would have to be to weak that it couldn't support the weight of the chassis.(

This is not all that bad an idea, You coukd actually use a guided cable assembly like used on lawn mower throttle controls, or bike derailer shift linkage. The outer cable is stiff enough and guides the inner actuator cable. The inner cable generate a·push or pull force, although the push is usually not as high as the pull. If you needed to use the stronger pull direction you could place the cable on the opposite side of the knee joint and use a lever pivoting at the knee to lock the knee straight. (see attached sketch).

Ted