ERP locimotion questions

[Joker94]

the following questions are directed at admin, but any one else that has had similar experience feel free to chime in.

Hi Admin

I have a few questions relating to your ERP robot.

First of all, do you find the HS-255BB servos strong enough for the things that you want to do plus some room for expansion of the robot? And how does the robots weight compare to the 4.8kg-cm the servos put out at 6v?

Secondly, how are you controlling the movement of the servos, do you have encoders for fine control or another means of tracking the servos movement?

Any response would be greatly appreciated.

Regards

Joker94

[Admin]

First of all, do you find the HS-255BB servos strong enough for the things that you want to do plus some room for expansion of the robot? And how does the robots weight compare to the 4.8kg-cm the servos put out at 6v?

I assume you are asking because you're considering building a similar robot. When I designed ERP, just like it is described in the tutorial, I didn't just decide it was 'strong enough'.

I actually sat down and determined the weights of things ERP will need to pick up, decided the exact speed I wanted him to travel, and calculated his exact weight in CAD.

Given that information, I used basic maths to calculate required torques of all the servos. This gives me the bare minimum torque I needed. Considering I had plans to upgrade ERP often, those torque requirements might change in the future. So I then multiplied that value by somethings called a 'safety of margin', which I arbitrarily decided to be 1.3 (1.3 x calculated torque = servo torque). And finally, I then went shopping around for servos with that torque. If I couldn't find an affordable servo with the torque I needed, I redesigned a bit, etc.

Many of the skills you need to do your own maths can be found here:
http://www.societyofrobots.com/mechanicsbasics.shtml
http://www.societyofrobots.com/calculator.shtml
http://www.societyofrobots.com/robot_arm_tutorial.shtml

Secondly, how are you controlling the movement of the servos, do you have encoders for fine control or another means of tracking the servos movement?

Its using open loop control, which means there is no position feedback. But that doesn't mean high error, because I've corrected for the error by experimenting before hand. I've been wanting to do what is called 'visual servoing', which means the camera is used to determine the arm location to pick up objects, but I've been too busy with other projects . . . Humans use visual servoing - our 'encoders' are way too inaccurate (close your eyes and tell me the exact angles of all 7 of your arm joints, and you'll know what I mean).

hope that helps . . .

[Joker94]

I assume you are asking because you're considering building a similar robot. When I designed ERP, just like it is described in the tutorial, I didn't just decide it was 'strong enough'. Tongue

Yes i am looking to building an experimental platform/butler bot that can be used for many different applications and experimentation, much like the reasoning behind ERP and a few other robot around.

The reason i ask the question was because I have heard many stories about people complaining about servos not living up to their expectations and not being able to control them as they would like. people have recommended that i steer clear of servos when it comes to robot locomotion. but then you look at ERP and its movement and accuracy, they look very much suitable and reliable to be used on such a platform.

Thanks for answering the questions. it is very much appreciated!

Regards

Joker94

[Admin]

The reason i ask the question was because I have heard many stories about people complaining about servos not living up to their expectations and not being able to control them as they would like. people have recommended that i steer clear of servos when it comes to robot locomotion.

Those people that gave you that bad advice probably didn't actually calculate anything, and only subjectively choose a servo too underpowered for their requirements.

As a mechanical engineer, my advice is to always design your robot 100% in CAD and do all the math. That way you don't waste time building something that wouldn't even work, and don't waste money on parts that aren't strong enough.

Oh one more thing, don't buy the HS-225's, they are complete garbage. The circuit board has a habit of literally going up in smoke and the gears break while still under spec:
http://www.societyofrobots.com/robotforum/index.php?topic=1300.msg23715#msg23715

[Joker94]

Oh one more thing, don't buy the HS-225's, they are complete garbage. The circuit board has a habit of literally going up in smoke and the gears break while still under spec:
http://www.societyofrobots.com/robotforum/index.php?topic=1300.msg23715#msg23715

thanks for the advice!

I have Auto Desk Inventor and CAD all of my projects before purchasing parts or building the chassis. I will definitely spend the time accurately CADing this robot and doing the related maths as i want this robot to be a long term project and something i can do a lot with. and will last!

by the way, brilliant job on the design of ERP.