Implementing Cobots: What They Don’t Tell You and How To Do It Right
All across the country in factories large and small, collaborative robots (Cobots) are sitting in unopened boxes, collecting dust instead of doing what they do best: enhancing productivity. I see this time and time again when I visit manufacturers that ask us for advice on Cobot automation.
Here’s the problem: Automation engineers are busy. They’re either too short on time to set up a Cobot or they quickly realize integration is more complex than anticipated. Successful integration requires more than the marketing pitch that Cobots are “easy to program.” Implementation usually requires interfacing to existing equipment, choosing or designing an end-effector (gripper) for the job, and building fixturing that wasn’t necessary when an employee was doing the job. Many automation engineers aren’t familiar with these aspects of Cobot implementation and get overwhelmed when asked to add them to their already demanding workload. Here’s how to ensure you don’t waste your Cobot investment:
The Cobot Is Only Part Of The Application
A Cobot is only one component of the full automation cell. For the physical setup, consider where it will be located in the work area. Can you roll the Cobot into the space where the employee formerly stood or must the work area first be redesigned? Once the Cobot is in place, does it hamper the employee from accessing the area? This is common because work areas were designed for people with high flexibility and dexterity: much more so than a typical robot. Typical Cobots (and industrial robots) have 6 joints, which makes them substantially less maneuverable than a human arm. Some Cobots are now available with seven axes (or joints), which gives them maneuverability similar to a human arm. In general, more axes equals more flexibility and versatility. Cobots with seven axes can fit into tighter spaces, reach around areas, and be located off to the side of the work area. Look for a model that won’t require the workspace to be redesigned to fit both the Cobot and an employee.
The next consideration is communication between the robot and the equipment with which the robot will work. If the Cobot will be tending a CNC or other machine, how will it interface with that machine? An older machine may not have interface capability or require a more complicated interface. Does your in-house engineer have the knowledge or expertise to configure that interface? Would it be easier to have the Cobot push the start button like an employee does rather than interfacing electrically with the machine? If so, can the Cobot reach the start button?
What Role Do Integrators Play In This Process?
An integrator will definitely get the job done, but beware of a few potential pitfalls.
Not all integrators are familiar with Cobots. Those who treat the project like an industrial robot can negate many of a Cobot’s benefits. For example: most Cobot applications, if designed correctly, do not require safety fencing, whereas most industrial robot applications do.
Then there’s the cost. Integration usually costs more — sometimes much more – than the Cobot itself. Does your supplier have the expertise to help you through the integration process? Will the supplier participate in the actual integration work or are you on your own once you receive your Cobot?
Many applications don’t require a full-scale integration program. A little coaching, a few tips, and maybe a little help with the accessories is all you need. Ask your Cobot supplier what help they can provide, and how much each service costs, before you decide to spend additional funds on an integrator.
How Is The Work Staged?
Most manufacturers don’t understand the need for and importance of staging the work the Cobot will handle, and it’s one reason implementations fail before they start.
A Cobot requires fixtures to be built that will hold parts in registration. This is a work-holding cost many potential customers don’t realize they’ll have to incur. Fixturing can be as simple as lining up the parts in a tray or as complex as a CNC-machined contraption with moving parts.
Without experience, it’s not always obvious to know the simplest solution. Ask your Cobot supplier to help. It will require a review or analysis of the parts the Cobot will be handling.
End-effectors: Another Additional Cost
End-effectors – grippers – are fundamental to an application’s success. What you use will depend entirely on the parts the Cobot is handling. If the Cobot is painting, polishing, sanding or deburring, that will require a different type of end effector.
End-effectors can be driven by air (pneumatic), mechanical, controlled electronically (these are the most flexible) or via vacuum with suction cups or other vacuum surfaces for picking up oddly shaped objects. Sometimes, for specialized functions, very complex devices must be custom-designed and manufactured.
You must consider the parts the Cobot will handle to choose the appropriate gripper. Ask your Cobot supplier what’s best whether or not the company can supply it. They must study your parts to make a recommendation. If they recommend a particular gripper without the benefit of studying your parts, find someone else to help with this.
Machine Vision: Should You Go There?
If Cobots had the eyesight and brain power of a 2-year-old, implementation would be much simpler. While no Cobot has that level of sophistication yet, a machine vision system can be added for an additional cost to relieve many of the challenges and requirements of fixturing. These systems have been around for a few decades, but there’s a catch.
Machine vision is notoriously complex to learn and program. Once programmed it can work very well, but the learning curve can be challenging. If you never plan to move your Cobot from job to job, a traditional 2D machine vision system can be a powerful addition, and most offer similar capabilities. 3D robot and Cobot vision systems are now starting to hit the market with varying results, which depend mostly on the specifics of the objects being handled. As with the other aspects of integration, ask your supplier if they sell vision systems and can set it up and program it for you.
A new generation of vision systems, such as our OB-Vision, are coming to market. They’re tailored specifically to guide the Cobot to pick up parts without the need for fixturing and don’t require programming. The vision system camera is integrated into the end of the robot arm and “looks” at the work area to find the parts to pick up. “Programming” this vision system involves “showing” the camera one sample part, and pushing a single button. While not as versatile as traditional 2D machine vision systems, this is a major benefit if you plan to move your Cobot or if you need it to work with a variety of parts.
Maintenance: Once the Cobot’s Up and Running
So now you’ve got the Cobot installed, the grippers designed and built, the interfacing done, programs written, accessories in place, and the job is running. You’re done – right?
Not so fast. In the real world, things change. Parts can vary a bit; components in the workspace move a little; something gets bumped, bent or tweaked. These everyday occurrences will often require that “easy-to-program” Cobot program to have it’s “program” adjusted.
Who’s going to do that? The Cobot supplier or integrator? Do you need the automation engineer to be on call, or is the Cobot so simple that its operator can make these changes?
Cobots that “learn” by being shown what to do are “taught”, rather than “programmed”. The benefit of this is that an engineer or programmer is not required to make basic changes to the robot’s “program”. Most production personnel can modify the robot’s path or movements easily.
Maintaining A Safe Operational Speed
A risk assessment is required for all Cobot applications. When an employee is going to be sharing the workspace, speed must be limited to 10 inches per second (250mm/sec). Therefore, if the application will be running around employees, you must limit Cobot speed to this “collaborative” level.
Of course, everyone wants faster-faster-faster. If you must limit speed (and this is the most common situation for a Cobot), make sure the Cobot’s controls can be locked so no one can change the speed. Make sure that your Cobot does not allow unauthorized or unintended speed changes without a password.
So you’ve made it this far. To quote Winston Churchill: “This is not the beginning of the end. This is perhaps the end of the beginning.” Cobots are the future of manufacturing and you will probably end up with many of them, so here are a few parting tips:
Do your homework before you choose. Talk to multiple Cobot suppliers and investigate multiple brands. Consider all aspects of the Cobot: from capabilities through implementation support to maintenance.
Find someone you can lean on for advice. Maybe you know another manufacturer who’s already begun implementation. Maybe your supplier has a lot of experience in implementations. Sometimes all you need is a few questions answered or a few tips. Someone who’s been there can save you a lot of time and money with a few simple conversations.
Start simple. Always, always, always start with the easiest job even if it’s not the one you need the most. Get the “easy” job running. Your team will learn the ropes; employees will become comfortable; and your Cobot will quickly start paying its way.