Hi,
I’m looking for a robot arm fast enough to throw a table tennis ball 1m into the air.
Are the lynxmotion arms capable of this?
um, with all due respect, pick some servos and do the math. servos have torque and velocity specifications available. personallyI’d start with a servo like the 5990 in a single axis and determine if you can do the job with that because if you find the length required is longer than the length of a stock arm you know right off the bat whether a stock arm with the beefiest servos will do the job or not. for a single axis, it is pretty basic physics math to determine the initial velocity required to throw a ball straigt up and go a distance of 1m before changing direction. now you can use the rotational velocity of a servo to calculate a minimum arm length to achieve that velocity, and then use F=ma to calculate the minimum torque required for the servo. that will define the minimum requirements for servos. in two dimensions it is more complicated as you need to introduce some kinematic equations to make the ball follow a linear trajectory, but even then you just need to determine the maximum velocities and torque at each axis to pick a servo. note this process is backwards from the single axis where it was simpler to pick a servo and see if it could do the job.
Am I correct that the “servo speed” from the specification only applies when the servo has no load?
Can I fit any servo onto robot arms like the AL5D?
Thanks for your math advice. I did the velocity calculation for a 20cm robot arm with a HS-5475MG. Assuming maximal speed, I get a speed of 1.09 m/s and a throw height of 6cm.
This can be done if you use a spring loaded catapult system and have a servo actuate a lever to release the arm. Some kind of winch motor could be used to reset the arm.
there is a fairly significant gear train in most servos so light loads will not significantly change the speed. when you start getting near the higher end of the torque range things like your power source and wiring will influence the maximum speed. that is pretty vague I know but the industry these servos come from isn’t so much about hard specifications as being competitive with the other guys marketing and making the servos as cheaply as possible without taking a black eye. my suggestion would be to make a torque calculation for the midpoint of the servo rating and be prepared to rework the servos with higher AWG wire and use a solid battery or beefy power supply. if you can get close with calculations at least it should limit the number of servo models to test with.
your estimate of 20cm seems fair for the AL5D arm. I know this suggestion is going to throw you into 2D calculations but perhaps additional velocity can be realized by flicking the wrist (?).
Thanks for your advice, now I got a much better picture of the situation!