RoboSavvy Designing and Fabricating a 1m tall Humanoid
by TiagoR
Tue Jan 08, 2013 7:19 pm
Hi all
Today we started a new variation of the hand.
Initially we took design inspiration from InMoov project and the beautiful mechatronics hand designed by Gael Langevin. However, his design involves every servo actuating the closing and opening of the hand. We had some issues in calibrating and repeatability in this type of actuation.
Instead of the fingers being actuated by the servos when opening and closing, they are just actuated when closing the hand. The hand's resting position is open by use of torsion springs in order to bring them back to open.
Here are some pictures of the hand prototype:
Hand Prototype
Comparing with real hand
Another deviation from InMoov is the length of the fingers. Since this hand is designed for accurate grasping and control and simulation, it seems like a better idea to have the fingers with equal length. We may get back to InMoov's beautiful exterior design features once all aspects of the hand, ROS control and simulation are working
Today we started a new variation of the hand.
Initially we took design inspiration from InMoov project and the beautiful mechatronics hand designed by Gael Langevin. However, his design involves every servo actuating the closing and opening of the hand. We had some issues in calibrating and repeatability in this type of actuation.
Instead of the fingers being actuated by the servos when opening and closing, they are just actuated when closing the hand. The hand's resting position is open by use of torsion springs in order to bring them back to open.
Here are some pictures of the hand prototype:
Hand Prototype
Comparing with real hand
Another deviation from InMoov is the length of the fingers. Since this hand is designed for accurate grasping and control and simulation, it seems like a better idea to have the fingers with equal length. We may get back to InMoov's beautiful exterior design features once all aspects of the hand, ROS control and simulation are working
TOSY Robot
by UncleBob
Fri Jan 11, 2013 1:59 pm
Can get some ideas from this Tosy robot from Vietnam featuring at CES.
http://www.tosy.com/en/r-d/topio-ping-pong/
120KG !!
http://www.tosy.com/en/r-d/topio-ping-pong/
120KG !!
by MarcoP
Fri Jan 11, 2013 7:56 pm
Here is a mega update on the hands:
The video pretty much speaks for itself so not much more to say.
Still not all the springs and mechanical stops implemented but enough to get a feel for how it works.
A bit more work and it should be good to go.
Already planning on how to add FSR for touch sensing when the mechanical aspect is resolved.
A closeup:
Regards
The video pretty much speaks for itself so not much more to say.
phpBB [media]
Still not all the springs and mechanical stops implemented but enough to get a feel for how it works.
A bit more work and it should be good to go.
Already planning on how to add FSR for touch sensing when the mechanical aspect is resolved.
A closeup:
Regards
by MarcoP
Fri Jan 18, 2013 12:13 pm
To add a sense of touch to the robot we are adding FSR's to the fingers.
THE FSR's have a connector with a thin transparent plastic where a conductive trace seems painted on and covered with protective varnish in the smaller version and a sort of tape in the larger version.
Since we want to place this in small fingers we want to get rid of the connectors. I did a test to check if soldering was possible to the conductive trace. Not a chance. The plastic layer melted away as soon as the soldering iron touches it.
The alternative was then to get electric conductive epoxy:
It's a two part component like normal 5min epoxy but it is mixed with silver particles which makes it conductive (and expensive).
The idea here is to cut the plastic strip, leaving just the sensor and glue very thin copper wires directly to the sensors:
Promising results. The epoxy seems as strong as normal epoxy and electric resistance is negligible.
We are using very thing (36AWG) teflon coated copper wire. Abrasion resistant and very flexible. The fingers were designed with channels to run these wires inseide. We manage to pass cables for 3 sensors in one finger and it still remained easy to flex.
Regards
THE FSR's have a connector with a thin transparent plastic where a conductive trace seems painted on and covered with protective varnish in the smaller version and a sort of tape in the larger version.
Since we want to place this in small fingers we want to get rid of the connectors. I did a test to check if soldering was possible to the conductive trace. Not a chance. The plastic layer melted away as soon as the soldering iron touches it.
The alternative was then to get electric conductive epoxy:
It's a two part component like normal 5min epoxy but it is mixed with silver particles which makes it conductive (and expensive).
The idea here is to cut the plastic strip, leaving just the sensor and glue very thin copper wires directly to the sensors:
Promising results. The epoxy seems as strong as normal epoxy and electric resistance is negligible.
We are using very thing (36AWG) teflon coated copper wire. Abrasion resistant and very flexible. The fingers were designed with channels to run these wires inseide. We manage to pass cables for 3 sensors in one finger and it still remained easy to flex.
Regards
by TiagoR
Fri Jan 18, 2013 5:09 pm
Hello all
Today I bring you some pictures of the mechanism that will allow rotation of the forearm.
This rotation will simulate the roll movement of the wrist, but it's applied near the elbow due to organization of the servos inside the forearm.
It consist basically on a 3D printed tube attached to the horn of a servo on one side and attached to the elbow metal frame on the other side.
It has to be a tube in order to allow cabling from the main controller to the forearm and power cables.
Arround that tube there is a bearing and the forearm is attached to the outside of the bearing.
The tube thats attached to the servo, the bearing and a part that keeps the tube tightened to the bearing are shown in the following picture.
The "counter" part to keep the forearm attached to the bearing is shown in the following pictures:
The "counter-forearm" build on the tube part:
The full mechanism is show in here:
It is not the full foream, just half disc to test if all the dimensions were correct, and if the mechanism works as expected.
Best regards
Today I bring you some pictures of the mechanism that will allow rotation of the forearm.
This rotation will simulate the roll movement of the wrist, but it's applied near the elbow due to organization of the servos inside the forearm.
It consist basically on a 3D printed tube attached to the horn of a servo on one side and attached to the elbow metal frame on the other side.
It has to be a tube in order to allow cabling from the main controller to the forearm and power cables.
Arround that tube there is a bearing and the forearm is attached to the outside of the bearing.
The tube thats attached to the servo, the bearing and a part that keeps the tube tightened to the bearing are shown in the following picture.
The "counter" part to keep the forearm attached to the bearing is shown in the following pictures:
The "counter-forearm" build on the tube part:
The full mechanism is show in here:
It is not the full foream, just half disc to test if all the dimensions were correct, and if the mechanism works as expected.
Best regards
by MarcoP
Sat Jan 19, 2013 11:10 pm
Hi billyzelsnack
The trick here is to use a standard bearing (http://www.bearingboys.co.uk/61706-2RS_Thin_Section_Ball_Bearing-82609-p) in an unusual application.
you get that illusion because we placed the screws very close to the inner and outer edge of the bearing.
Got a cross section for you:
Red is the bearing and blue is open space. The stationary part grips the inner ring and the rotating part grips the outer ring. (Didn't have the parts quite aligned so the top portion of the screw hole does not show up).
Regards
The trick here is to use a standard bearing (http://www.bearingboys.co.uk/61706-2RS_Thin_Section_Ball_Bearing-82609-p) in an unusual application.
you get that illusion because we placed the screws very close to the inner and outer edge of the bearing.
Got a cross section for you:
Red is the bearing and blue is open space. The stationary part grips the inner ring and the rotating part grips the outer ring. (Didn't have the parts quite aligned so the top portion of the screw hole does not show up).
Regards
by billyzelsnack
Sun Jan 20, 2013 1:04 am
I see now. Very cool. Kinda reminds me of a Freestyle BMX detangler.
http://www.jensonusa.com/Odyssey-GYRO3-Detangler
http://www.jensonusa.com/Odyssey-GYRO3-Detangler
by MarcoP
Fri Feb 01, 2013 7:06 pm
Hi guys
Another large update on the hand status.
First try of fully assembled system.
The system is being controlled manually with a Arduino compatible board .
Still some adjustments to me made in finger angle and rubber like texture to grip objects but still a very nice functionality.
Regards
Another large update on the hand status.
phpBB [media]
First try of fully assembled system.
The system is being controlled manually with a Arduino compatible board .
Still some adjustments to me made in finger angle and rubber like texture to grip objects but still a very nice functionality.
Regards