Robot Arm

[Tarek]

Hello everybody,
I'm newcomer to this great forum! I'm just another person who is interested in Robotics and would like to get started as soon as possible. I'd like to make a robot arm.
How could I get started? Please tell me everything from the very beginning and give me the suitable links that will help me to learn more about making robotic arm.
I'm planning to buy a book for this purpose. Which book would be better? Please do reply.
I look forward to hearing from you all

[Admin]

Actually, a few others have asked the same question too . . .

here are the posts:

http://www.societyofrobots.com/robotforum/index.php?topic=34.0

http://www.societyofrobots.com/robotforum/index.php?topic=68.0

hope that helps!

[Tarek]

Hey Admin,
Thanks for reply as well as the links. The links are great!
Actually, I'm gonna make a robot arm that will be used in arc welding purpose - a robotic arm that will hold the manual welding torch and manipulate it according to program. I'm mechanical engineering student and I'm gonna do this as my undergraduate project.
Will it be very tough project? How can I make this project easier? What should be the lengths of the links for this purpose? How strong servo will I need for this?
Thanks

[Admin]

Hmmm its doable. There are a lot of books out there for robotic welding.
The affordable automation book on this page has a lot on industrial robotics:
http://www.societyofrobots.com/robottheory.shtml

Key points I remember from them:

• Welding requires very strong arms (to carry the welding device, and hold a high precision position)
• Bright light blinds any cameras, making sensors useless
• The robot arm end effector needs to be highly resistant to heat
• People have been killed by industrial robots, safety is key

Instead of welding, how about brazing instead?
http://www.societyofrobots.com/mechanics_brazing.shtml

I think that brazing instead of welding will simplify your project greatly without sacrifice in quality. It will allow you to shrink your robot to a much smaller, much easier to build, much cheaper device.

As for your other questions:
Shorter link lengths are better in that they

• bend less, allowing higher accuracy and less 'flopping around'
• require less torque to move, and hence smaller/cheaper motors

Shorter link lengths are worse in that they

• have a shorter reach
• can only work with smaller parts

To calculate the torque your servos need:
http://www.societyofrobots.com/mechanics_statics.shtml
If you weld and not braze, you will need motors much stronger than servos.

[Tarek]

thanks for wonderful suggestion! splendid....i must say!!
I realize brazing instead of welding will be easier to build. But the problem is my supervisor is restricted to welding as he thinks the use of robot for welding purpose is more common and has significant positive impact in Industries. However, I keep trying to convince him to make it brazing instead of welding.

What if I consider Lynx 5 robotic arm for welding? I think it has strong arm and stout gripper for precision. But I don't know about sensors-whether the sensors become useless during welding due to bright light or not.
What do you think? Will it match with that one which I attempt to make for welding? Please check it out:
http://www.active-robots.com/products/robots/lynx5-details.shtml

Lynx 5 Robotic Arm Specs
No of axis = 4 + Gripper
Servo motion control = local closed loop
Height (arm parked) = 5.5"
Height (reaching up) = 14.5"
Reach (forward) = 11"
Gripper opening = 2"
Lift weight (arm extended) = approx. 3 oz
Weight (without batteries) = 18 oz
Range of motion per axis = 180 degrees
Accuracy of motion per axis = Servo controller dependant (SSC32=.09 degrees)
Servo voltage = 6 vdc

Any additional suggestion will be highly appreciated.
Thank you once again

[Admin]

Hmmm but welding robots have been around since the 70's . . . the technology is way developed by now Im sure . . . what kind of research does it plan to do with it, can i ask?

As for the lynx arm, its way too weak for welding (look up the weight of the welding equipment you plan to use). You will need an industrial strength arm:
http://www.robots.com/industrial-robot-arm.htm
(look at the picture links on the right for the most popular robot arms in use today)

Typical costs of an industrial robot arm is like $4k per degree of freedom. You will need at least 3 DOF for welding, 4 being recommended.

[Tarek]

Acutally I don't need to make that much strong and precise robotic arm welder that are now using in Industires.
Have a look at this robot welder: http://www.fronius.com/welding.technology/products/robot/
this seems pretty much simplier than your those gigantic industrial arms.

The robot arm that I tend to fabricate during my project will use manual arc welding equipment. While some might think that a robotic GMAW torch is nothing more than a manual gun attached to a robot arm, there are significant differences.

Take a look at this Figure "the basic arc welding circuit" :
http://www.lincolnelectric.com/knowledge/articles/content/arcweldfund.asp
My goal is simple: my robot arm will hold the electrode holder that is demonstrated in the figure and manipulate the electrode holder in the way sothat the tips squeeze together the two pieces of metal that need to be welded.

By the way, I'm also considering brazing as I really appreciate your suggestion. Would you please give me some links that demonstrate robotic arm for brazing purpose?
Thanks

[Admin]

Hmmmm that welding robot you found looks very affordable and just what you need. I am kind of curious of the motors (torque) they use for it . . . 
Also curious of the exact price . . . Considered getting a price quote?

Do you know your torque and holding force requirements to use your welding equipment on a robot arm? You can call them and say 'i am interested in buying your robot arm, but i have questions - what are the capable torques on these certain joints?' that could give you an idea of what torques you will need . . .

So I thought about the brazing problem some more . . . you will need a feeder for the filler metal, and also apply flux beforehand . . . it might not be as simple as I thought before . . . I am starting to think that your welding might be easier in terms of number of processes required. I still argue brazing is cheaper, lighter, and requires less precision pathfinding 
But I think you should stick to welding afterall . . .

A robot brazing arm:
http://homepages.cae.wisc.edu/~dongshin/resume/index.html
(scroll about halfway until you see robot arms)
But they are big enough for brazing cars . . . And surely too expensive . . .
Robot brazing arms are in the category 'brazing and soldering' if you look around for them.

[Tarek]

So you are telling me to stick to welding . Thank you so much!Well.....actually, this is my school project and my budget is limited. So I look forward to making it as simple as possible.
I sent them an email asking the question that you told me to ask.......but have not recieved any reply yet.......still waiting for reply.
In the meanwhile, do you have any other suggestion? What other things do I need to know before starting this project?
Thanks

[Admin]

Hmmmm I think you know enough now to get started. As you start designing you will find things you dont know and will have more questions. Just ask then 

Read up on kinematics (and inverse kinematics) and determine your required torques of each joint. If you get that done, your in good shape . . .

[Tarek]

I haven't recieved any reply yet........and never thought of getting any response from them

Anyway, the height of electrode would be 10 inch. So I need additional 5-6 inch height for arm's movement. The desired movement will be rectangular movement (in X-Y direction) as illustrated below:
  <-- <-- <-- <-- ^
  l                      l
   --> --> --> -->
  ==========

The numbers I assumed are as follows:
Mass of first link (joint with base) = 600 g
Mass of each of other links = 300 g
Mass of each servo = 150 g
Mass of gripper = 300 g
Mass of electrode holder = 600 g
L1 is 40 cm and L2 is 20 cm and L3 is 5 cm.
Here, L1, L2, and L3 are the links

After calculation, I got:
(.6*40/2+.3*(20/2+40)+.3*(5/2+20+40)+.15*40+.15*60+.3*65+.6*65) =119.25 kg*cm or 12.16 N*m required

119.25 (kg cm) = 1656.07 inch ounces

Now what do you think? What kind of servo do I need for this?
Thanks

[Admin]

So you will be operating in only the X-Y plane?
Consider using an X-Y cartesian robot instead of a full arm (see image). This is the same as those etch-o-sketch's from back in the 80's.

You could also consider keeping the welding equipment fixed and motionless, while instead moving the part to be welded by a robotic arm or cartesian robot arm . . . Assuming moving the part is easier than moving the welding equipment.

If you want 1600oz+ torque:
http://www.servocity.com/html/robotzone_servos.html
They claim '1,715 in-oz. of Power'
Its the strongest I can find and will cost $100+ (if you include servo too) . . .

[Tarek]

well thanks……..hmmm keeping the welding equipment fixed and motionless, while instead moving the part to be welded by a robotic arm would be easier.........but perhaps it may not be the correct method of welding that is being done usually …… and I’m afraid it may change the main objective of my project.

I've made some amendments in my calculation.
L1 is 30 cm and L2 is 30 cm and L3 is no need.
Here, L1, L2, and L3 are the links
base height + L1 = 40 cm (I forgot about base during my previous calculation)
L3 is not necessary because electrode has it's own height.....so the attachment of only gripper at the end of the link L2 would be fine.

So my assumptions are:
Mass of first link (joint with base) = 400 g
Mass of other link = 400 g
Mass of each servo = 150 g
Mass of gripper = 300 g
Mass of electrode holder = 600 g

And now my calculation stands:
(.4*30/2+.4*(30/2+30)+.15*30+.15*60+.3*60+.6*60) = 91.5 kg*cm or 9.33 N*m required
91.5 kg cm = 1270.71 in-oz
Is my calculation okay now? If okay, then robotzone's '1,715 in-oz. of Power' would be more than enough.
And obviously gearing is essential here as far as high torque is concerned.

Would the required torque be total torque or torque that requires from each servo? If it is total torque, then i think HS-645MG + 5:1 ServoPower Gearbox = $79.90 would be enough……they claim 665 in-oz. of power.
So 665*2=1330 in-oz which is greater than my required torque (1270.71 in-oz)
http://www.servocity.com/html/hitec_servos.html
What do you think about this?

[Admin]

So your torque calculation for base joint is correct. But I think your third link is going to have some set length greater than zero since the last servo, the gripper, and the electrode cant all be at distance zero.

You actually need to calculate torque required by each joint to know what servo you need at each joint. Thats easy though because it basically uses the same equation as the base joint. The base joint is always required to be the strongest so the others should be much easier to do.

You might also want to calculate speed of your arm. Welding I am sure requires some optimal transversal speed, so you need to see if the servos can move your arm about that speed. This is a somewhat complicated calculation in that it depends on the speeds and link lengths of all joints combined . . . but if your base joint can do the correct speed alone then you are fine.

[wolfy_9005]

A good "claw" is a "boss head", used in chemistry which you stick on retoert stands to hold test tubes and the like. Just stick a motor on the screw that opens/closes the claw and u can use it to pick stuff up and tighten it as u want.