Pivot Joint Mechanism
Saikumar Vasu
I am trying to build a 5 axis Robot Arm. I am using 6 Hitec 1/4 scale
servos for the joints which make the robot arm heavy by itself. The reason
why I went in for the 1/4 scale servo is increased torque requirement for my
project. I am not good at Mechanical designs as I am a computer science
related person.
My question is how can I construct a pivot joint for the base rotation which
is sturdy enough for a holding a weight of the 6 1/4 scale servos and 1/4
kg of weight in the grasper ( end-effector ). I know I need some sort of a
sprocket and chain mechanism but dont know the complete details and the
complete parts required to assemble such a design.
Any help would be appreciated.
Thanks a lot.
BTW, Happy New Year.
--Sai.
Gordon McComb
Better than sproket and chain (or belt and pulley) for some of the
linkages may be a rigid link. I made an arm using a combination of
1/4-scale and standard servos. I'm not sure you need 1/4-scale servos
for the write flexion and rotation (and whatever you're using for the
gripper closure). If you do, for whatever torque that's required, then
1/4-scale servos may not be enough for the elbow and shoulder portions.
Likely too much weight.
When you say "base rotation" it sounds like you're building a polar
coordidate arm. My prototype, a revolute (human-style) arm, can easily
lift four ounces (it was made to lift a 12 ounce can of soda), and it
uses the following:
1. Nema 23 stepper for the shoulder rotation. (There is no shoulder
flexion in my model. The shoulder motor is actually part of the body of
the robot. I use a 6" expansion spring to balance the weight of the arm,
so that the stepper only has to move a few ounces worth with lifting the
arm. This is key to reducing the need for a higher torque shoulder
motor.)
2. 1/4-scale servo for the elbow flexion. (I used 1/4-inch Delrin
sprocket and chain here, which is light-weight. You can sometimes find
this surplus, but it may be necessary for you to look at Small Parts or
similar prime source. This motor is attached behind the shulder rotation
pivot, so its weight acts as a kind of counterbalance. As such, its
weight does not "count" for the lifting force required of the arm.)
3. 1/4-scale servo for wrist rotation. (This is the one connected to the
rigid link. The link is a length of servo rod, the same kind used by RC
modellers. It is attached on both ends using standard RC servo rod
connectors.)
4. Standard sport servo for wrist flexion.
5. Standard sport servo for gripper closure. (This is operated by means
of small gears, and is actually the most complicated part of the
arm...as grippers often are.)
I built my prototype about 10 years ago, and I used Super Glue for some
of the parts. I didn't know then that Super Glue, especially over smooth
surfaces, lacks long-term holding power. So, the end-effector has
broken off over time, but if need it, I can post pictures of the art at
my site. Though not a polar coordinate design, some of the concepts may
help. It uses only standard hardware you can find at the hardware store
and RC modeling shop.
(BTW, the above arm is NOT the only from either edition of my book. It's
a different one, and alas, harder to build, requiring some shop
experience. But it's overall a better design.)
-- Gordon
Robot Builder's Bonanza, Second Edition
http://www.robotoid.com/
Saikumar Vasu
I don't know whether you got my previous mail which I sent to you. So I
am reposting it .
Thanks,
--Sai.
----------------------------------------------------------------------------
-
Gordon,
Thanks for the tips you had provided. It was really helpful. I own a
copy of your book ( 2nd edition ) and it's really fantastic source of
information.
The design I had followed was a rigid link for the elbow as you have
suggested
For the elbow and shoulder I used two 1/4 scale servos and for the wrist and
the gripper I used standard servos.
I don't know whether the design I had followed was good enough, but just
learning from mistakes.
I am still skeptical on the type of mechanical joint which will give a
stable and sturdy base rotation as I had drawn in the picture attached. I
don't think a rigid link is good for the base as the whole weight of the
robot will rest on the servo axle.
Could you please send me /post some pictures of your robot.
Are there any books which shows designs of some of the industrial robots
and the type of mechanism they use?
Thanks for you help,
Wish you a very happy New Year.
--Sai.
----------------------------------------------------------------------------
--
"Gordon McComb" <gmc...@gmccomb.com> wrote in message
news:3C30BD...@gmccomb.com...
Gordon McComb
>
> Hi Gordon,
> I don't know whether you got my previous mail which I sent to you. So I
> am reposting it .
> Thanks,
Yes, I got the e-mail, and was needing to wait until the weekend to take
some snaps of my arm to send to you. This board doesn't accept binaries
(well, it'll technically take them, but it's not a binaries newsgroup),
so I'll post them on robotoid.com when they're done.
For the base, I'd really recommend something with a good thrust bearing
shaft, like a NEMA 23 size stepper. The 1/4-scale servo might work, but
they are not made for a "turntable" type of mechanical arrangement. They
also only have about 180 degree (plus or minus a few degrees) of
rotation, and for your arm design, you'll want 360. There *are* some
larger servos for things like gear retraction that are 270-720 degrees,
but they get fairly costly.
Finally, the work envelope of a stationary revolute arm "missses"
portions. A polar arm may be better, and they're easier to build. Since
you have RBB, the plans for such an arm can be found in Chapter 26. If
you want to build it, try the following changes/additions:
1. Use extruded aluminum instead of steel shelving standards. Not as
strong, but you don't need strength here.
2. Add a cross-piece to the base to support the other two corners of
the turntable. It greatly enhanced stability.
3. I made the center post and "elbow" shaft using hardware I'd know
anyone could get (like closet pole holders for spacers). Feel free to
substitute if you find other/better hardware.
Alas, no servos (only steppers) but my prototype arm could lift about a
pound. And I realize a stationary revolute arm just looks a lot more
cool than a polar arm.