Thread: looking to get into Robotics

hey everyone, Im Alex, and Im an apprentice millwright, and I am looking to get into robotics. Can anyone give me some tips and pointers as to where to start and what to look into and expect? thanks

Originally Posted by alsm6887

hey everyone, Im Alex, and Im an apprentice millwright, and I am looking to get into robotics. Can anyone give me some tips and pointers as to where to start and what to look into and expect? thanks

You must say exactly what you need. Regards Libor

Is this something you want to do where you work for a specific task or a general statement? Do you want to get into the business of installing robotics or just know enough to select one for a specific task?

Most of the millwrights that I know are weak when it comes to getting behind a laptop/programming and understanding electricity/controls such that they can work their way around a machine and troubleshoot proxes, light curtains, drives, controllers, networks, etc.

A good robot systems integrator has all of these skills plus understanding of end effectors, kinematics, coordination with conveyors, etc.

As Libor said - Say exactly what you need (and want to accomplish) with said robot.

Proficiency in PLCs and ladder logic would be an important prerequisite.

Just to add, experience in C++ programming would also be a good thing to have.

And perhaps an up-to-date and current will, in case the robot bitch-slaps you into the adjacent cell.

What do you mean?
Repairing?
Installing?
Programming?
Specifying?
End effector design?
Application engineering?

It's kind of a big field.
If you 'got into robotics' what is it you would like to tell some one you know how to do?

If we hear back from the OP...This is sort of one of those areas where there's almost "too much to know". Definitely agree if you can be a solid electrical or mechanical person to start and then reach across into the other discipline as much as possible, you can be really valuable.

So me, I'm not a millwright, but a mechanical person at heart. I can work safely with robots, touch up their programs, have a basic understanding of what the programs are doing and how they interface with PLCs/safeties/their own I/O -- and how to recover them back to an automatic cycle, I have made many repairs to their end effectors. I have replaced the "shoulder" bearings in robots where the original grease pack failed, rolling elements fell out as dust, and the robot was clunking along like a tin man and wouldn't hold position, and to boot, the robot was inside such a maze of other automation the total cost to simply replace the robot would have been astronomical and risky. I have also repaired broken bolts (M30s) which hold down the robot bases, I have replaced failed axis motors. I'm not trying to trumpet my virtues here, just trying to illustrate the above is basically scratching the very outer surface of what a robot does.

I would be at a loss and have to start a bigger learning curve if I had to develop brand new programs, use the robot program to make decisions with internal logic, setup mathematical offsets for repeating part arrays, or do any major reconfiguration of I/O. If I had to interface with a new process such as vision measurement, or integrate it with a VMC or a welder, etc, I'd be at a loss.

In other words, can you be a millwright and be familiar, useful, and helpful with a robot? Yes. If you want to be a robot expert my free suggestion is you need to be an apprentice electrician or controls engineer.

sorry for not getting back sooner guys, but thanks for the responses. I saw a lot of people were wanting to know why i want to get into robotics and what I want out of it, so I will do my best to explain my intentions.
Robotics have always been an interest to me, as was milwrighting, and i have always found robots to be very interesting, and challenging, and i like a good challenge. I am very interested in pursuing the whole deal, instalation, repairs, programming, etc. being good with both mechanics and computers, it only seems natural for me to enter into a scene where these two areas overlap, and thats robots. As a millwright, my college course required me to have a good understanding of electrical and pneumatic systems, as well as hydraulics. I was good in all three areas, and hopefully this will help?I want to be able to do the course for robotics and say that I am trained in robotics, and be able to go to any major factory or production plant, look at the robots in oporation, and be able to know exactly how and why they work, and repair a damaged system.

Originally Posted by Radar987

Proficiency in PLCs and ladder logic would be an important prerequisite.

I really want to learn this myself. What are some ways? I am a mechanical engineer/cnc machinist by trade.

There's books on ladder programing, but if you want to get into robotics, so what I did. Go buy a decent used robot, teach yourself how to install it, master it and get it up and running. Then you can experiemtnto see what you can do with it. It's how I learned to run a VMC, a horizontal, my lathe, My Integrex, and a pair of smart turns with a gantry robot. When you break it down into logical steps, it's pretty basic for any of these machines. I'm currfently working on 3D milling with an 8 axis robotic system. Turns out this is a bit bigger nut to crack than anticipated, but I've learned a ton of other tricks along the way, so it's all worth it.

thank you...nice post

Now just to make sure I am following, Ladder logic is a system whereby the robot or PLC must complete one action before moving on to the next right? Could I compare ladder logic to a ladder diagram for an electrical system? Also, is there a specific company or website etc that I could access to get information on robotics so I am somewhat competant when I start out? As I have said previously, it has been a personal dream to be able to install, maintain and program robots, so obviously I am very serious about this, and all input would be very valuable to me!

I'm currfently working on 3D milling with an 8 axis robotic system.

I'd like to see the thread that follows that project....

Anyhow, since you seem to be starting from scratch, it is my suggestion you go to half.com (or a place like it) and get a book on ladder logic to get you started in controls. Then you can find a bunch of info on robots just by Google searches. For instance, getting to know the different coordinate systems (world, tool, joint, user are some examples) of a robot would be considered a basic skill. If you are competent in any kind of programming language it will go a long way as syntax will be different but the logic carries through. Also, if you plan to troubleshoot and repair, studying how a cnc system uses servos, amplifiers, encoders and axis cards for position control is a must. I could probably go on all night. Just study anything that has to do with industrial controls. You could be taught how to jog a robot in minutes, but in order to have a complete understanding you got to get close to em and dig in!

I mentioned PLCs as usually robots are used in conjunction with other automation, so at some point there has to be a master coordinator between the various elements. PLCs are often this coordinator. However the robot has some of its own I/O capability and so it can be the coordinator for a smaller set of external processes.

For example, if you had a mig welding robot, the robot controller (Fanuc is what I'm familiar with) can signal the welder to do its thing. However, if you are into controlling and coordinating multiple robots in a single cell, along with other automation, clamps, conveyors, etc, then usually a PLC is needed.

The Fanucs I'm familiar with have their own programming language which is more English based. PLCs use symbolic ladder logic (with the useful descriptors! ) but are really powerful these days with all sorts of functions that go way beyond simple decisions of input sensors and outputs.

Anyway...best way to learn is to get a job in a plant that uses them and work with mentors on industrial maintenance (firefighting and troubleshooting). Also could go with machine builders/cell integrators as those people setup, debug and warranty their systems.

You could also go to school. The robot mfgs offer classes (for a fee). You can download the Kawasaki course catalog here: Kawasaki Robotics - Training - I'm sure the other mfgs offer the same type of thing.

PS: The current controller model for Kawasaki is the E, previous (up to last year) was the D. They are very similar with the E having even more functionality than the D.

Tony,

Why would you need C++ experience for robotics and automation programming, unless it is used to collect data on a plant wide basis and to integrate with industrial strength databases such as Oracle or DB2. Just curious, seems like overkill for the OPs initial needs.

Thanks

Von

Originally Posted by Tonytn36

Just to add, experience in C++ programming would also be a good thing to have.

Von,

The AS Language (Advanced programming language) for Kawasaki (and as I understand the advanced language for several other controllers also) is based on the C++ programming language. Most C++ will work with the Kaw controller. I don't program much C++, but it helps a lot if you understand the syntax.

The basic "block" programming most controllers offer is just that.... very basic. You can't do a whole lot with it. The Advanced languages of the controllers have tons more functionality. If you are just doing very basic tasks with very little interaction with the outside world block programming is fine. This is not usually the case in automated systems. The advanced languages allow you intimate control over the robot, the path, the I/O, program execution flow, etc.

You didn't really say how much you wanted to invest. If you've got a spare $10k, I'd go buy a Fanuc 420iF and start playing with it. If you have $100 I buy soemthing like a mini sumo robot that uses a parallax basic stamp. Silly as it sounds, If you can program one fo those, you can fairly readily transfer the simple structured programing you'll learn in P basic to a real robot.

You need to learn how to pull an output high or low to casue an action, and how to read an input and use it in your program. Mastering the basics of I/O is essential. Once you learn a basic stamp, you can move to most other micro controllers like an arduino, etc. Working with I/O in a Fanuc robot is even easier, since it's simple to invert that staus of an I/O point, or output most any signal you want. Setting up complimentary I/Os is a simple task.

Once you hae the basics, you add some pneumatic grippers mounted to a wall. Add another table with an array laid out in felt pen and practise picking up blocks (pretend its raw stock) from the table and putting it in a pneumatic gripper (pretend it's a chuck). Take the part out, flip it and replace it just like you might need to complete 2 operations on a lathe. Have the robot remove the finished part and stick it on the table and repeat through your array. When you decide to hook it up to a real machine (or 3) it's not much harder. The consequences of making a mistake just increase.

If you can afford something with collision guard, get it. Anyone that claims to have never slammed a robot into something either hasn't used one much or is lying. You will crash your first robot. You just want to minimze the damage it will cause when it happens. The first time you watch a robot fold over a 1/4" thick A36 plate like it's butter, and keep right on working will be an eye opener, and you'll make sure to stay out of the robots kill zone (hopefully).

Robots are cool, capale and a lot of fun. Complex 3D toolpaths are a PITA (for now, but it's getting better). I see now why I have yet to see a milling robot operating other than on a video. It sure as heck isn't as simple as a CNC machine. For pick and place applications it's simple as making a pie.

I thought I might add that learning how to repair CNCs can only help you as far as getting your foot in the door.