Using linear actuators on a robot arm

I see your idea for this.  The only problem would be getting the actuator to
send some feed back like in a servo.   Then the controler could judge the
distance the actuator has extended.  Another great idea would be to get some
feed back on current draw as well then the controler could judge the weight
or load on the actuator.   At the very least you could put in some bump/stop
switches to tell if the actuator has fully extended or fully retracted.
Anybody else putting more thought into this...

I know I just posted, but here's something I found right away.  Take a look.

http://www.tritonlinear.com/tritoncomspecificationsenglish.htm

Seems to be just what your looking for.

Here's something a little more RC/Robot  friendly...

http://www.emsjomar.com/linear.aspx

Good Luck.

Aren't the devices you are talking about nothing more than a screw-drive
system driven by a motor with a forward/reverse switch?  It gets the power
simply by using gears and sacrificing speed for force.  If the parameters
are what you are looking for, then it should work fine.  But as others have
said, you'll probably want to add position feedback (simplest case just
limit switches).  There are plenty of options of building your own system
of gears and or levers with standard motors as well to get whatever
leverage you need - including a screw drive - or pulleys and cables for
example.

I think you don't tend to find those types of linear actuators in robots as
much because they tend to be too slow for many applications. But if high
force and low speed works for you I think they would be a fine option.

--
Curt Welch                                            http://CurtWelch.Com/
curt@kcwc.com                                        http://NewsReader.Com/

If there are gears in the system wouldn't it be easier and cheaper just
to in stall a Gear Tooth Counter? Just a thought.

Casey

if your into building things here are crude images of a satillite sized
actualtor I made

http://eds.dyndns.org:81/~ircjunk/actuator/

as for positional feedback, something like a potentiometer at the joint
may siut you fine.

  dan

slow_mac_modem@yahoo.com wrote:

I'm dealing with this problem too, in trying to make a robot arm.
Servos just aren't strong enough, even the expensive ones. As far as
getting position feedback on the arm, I would suggest measuring the
actual position of the output arm rather than trying to measure
something in the geartrain and calculate the position. If your
home-made parts are in the same range of precision mine are (which is
to say, poor), then it's hard to get good feedback by calculating where
the arm *should* be. A potentiometer on the axis of rotation works, or
you might go for a resistive "bend sensor," as seen in those old
Nintendo Power Gloves.