In 2017, in response to a Boston Dynamics video, billionaire and pseudo-engineer Elon Musk infamously tweeted, “This is nothing. In a few years, that bot will move so fast you’ll need a strobe light to see it. Sweet dreams….” Whether or not Musk’s ominous prediction comes true for Atlas (the robot in the video), he raises a good question. How critical is high speed in robotics?
Of course, when it comes to industrial robots, this question can mean close to nothing unless we define what type of robot we are concerned with and what application the robot is performing. For example, speed considerations for a six-axis automotive paint robot are quite different from those of a delta robot performing assembly in an electronics factory. Applications such as painting and arc welding have maximum speeds due to the processes. A MIG welder can only lay down so many inches of bead per minute. A paint robot prioritizes smooth, sweeping motions over fast, jerky joint movements while spraying, for obvious reasons. But what about the many applications, such as pick-and-place or assembly, where there seems to be no ceiling to movement speed? Why don’t manufacturers choose robots that move so fast you can barely see them work?
Larry Allingham set out to answer this question in a recent white paper from ATS Automation. According to Allingham, historically, a high-speed assembly robot would not have made sense in cases where high output was not required. However, the technology of today’s robots makes the question more complicated. Allingham details three cases in which a high-speed assembly robot could make investment sense for manufacturers, even without a high output requirement.
- Multi-product versatility: A versatile machine, such as a multi-axis robot rather than a purpose-built automated assembly machine, can run another product when it isn’t being utilized for the original product, whereas a purpose-built machine will simply collect dust. This higher utilization leads to better ROI.
- Consolidated operations: If your production includes two machines running two separate tasks at x speed, it might be possible to install one robot with a speed of >2x. This robot would then be able to complete both tasks concurrently with the same output rate, consolidating your operations.
- Pre-engineered machines with high-speed capability: If a pre-engineered solution provides just what you need, and it also comes with high-speed capability, the speed is just a bonus.
What Types of Robots Are Used for High-Speed Assembly?
The type, model and features of the robot you’ll need for a specific assembly application depend on the reach and payload required. For example, small and light assembly of electronics could be done by a delta or SCARA (selective compliance articulated robot arm) robot. These robots are typically limited by payload but can also be extremely fast. For example, the four-armed Omron Adept Quattro delta robot has a maximum working area of 1500 mm, a payload of 15 kg, and can perform 300 picks per minute.
SCARA Robots for Assembly
Most SCARA robots have a two-jointed arm that can move the position of the end effector in x- and y-axes, with a linear z-axis, as well as a rotational axis. The SCARA design was originally created for assembly. The robots are capable of high accuracy and high speed. However, their design limits them from replicating pitch and yaw motion—so in certain applications, such as following a 3D curve, a SCARA robot may not be suitable and a six-axis arm robot may be needed instead.
Cartesian Robots for Assembly
The form factor of a cartesian robot is likely best known today as the system by which many 3D printers move the print head in straight lines along the x-, y- and z-axes. For assembly, additional rotational axes may be added.
Cartesian robots rely on linear actuators, such as a belt or ball screw, to provide motion. This makes them relatively inexpensive and modular. Cartesian robots are ideal for work that requires a large, rectangular work envelope. Whereas SCARA robots typically have a maximum reach of 1000 mm, cartesian robots can be designed to reach 5000 mm or more.
While cartesian robots are sometimes custom-built for purpose, building and programming, these robots can present a complex engineering task, both in terms of the kinematics and the specifications of the motors and frame for the forces involved in applications.
What About Low Speed?
As Larry Allingham pointed out, if a pre-engineered solution provides value for your project, the high-speed capability that comes with it can just be considered a bonus. The question could be asked, what about the flip side of that coin?
According to the three cases Allingham detailed, multi-product versatility can allow you to increase utilization and achieve better ROI. For this reason, in some cases, a collaborative robot could be a solution for assembly in high mix, low volume applications, since collaborative robots are by definition more user-friendly to program, teach and redeploy—despite their low speed.
Design for Assembly
Another consideration for high-speed automated assembly is the product’s design. The fastest robot on the market won’t help you if its speed is wasted on a complex or unreliable assembly process. Design for automation may be worth the investment if it makes the task faster and more reliable for your robot.
For example, a product that has been previously assembled by human workers may require the worker to hold parts together with one hand, then operate a screwdriver with the other. A human worker can also turn parts over, working on multiple sides of an object. Holding two objects isn’t possible for most robots, which have one end effector, and handling objects or changing tools adds time to an assembly process. For automation, a better design might be to have the assembly built up in layers while sitting on a surface or being held in a fixture.
Parts designed for automation will often snap together, avoiding the need for fasteners. In addition, parts that are self-locating, such as a pin with a chamfer to allow it to bump into place if it contacts the edge of a hole as it is inserted, will help prevent faults and improve repeatability.
Another consideration is the quality of parts used in the assembly process. For example, a human worker can take a second to remove a bit of mold flash or a burr before assembling a part, but a robot may crash if a defect gets in the way of the programmed path.
Depending on the assembly task, force sensing may be used to prevent the robot from damaging parts as it moves, such as driving a screw without cross-threading or stripping it. Force sensing may require slower motion, so other mechanical means of preventing these defects may improve cycle time, such as a fastener or tool that cams out when the screw is in place.
Lastly, the way parts are fed to the robot can greatly affect assembly speed. If the robot needs to pick parts from a bin using a vision system, that can add seconds or minutes to the cycle time (but may save time in other areas of production). If parts are presented repeatably in a fixture so that the robot can simply follow a path, the robot can work much faster and more reliably. Avoiding the need to reposition or manipulate parts will also save time.
Quality and Output vs. Cycle Time
Ultimately, the goal of high-speed automation is not speed; it’s high output. Like the tortoise and the hare, a reliable and consistent robotic cell can easily outperform a high-speed robotic cell if the high-speed cell breaks down or faults more often, requiring a worker to go in and reset it. In addition, if a slower cell will produce better quality output and fewer defective or no-go parts, that increases output too. Just like you may need to lower the speed of a CNC milling machine to produce the best quality part, the same may be true for your assembly robot.
Even if you don’t need a strobe light to see your robots moving, the best solution is one that works reliably and doesn’t have you tearing your hair out or the plant manager swearing off robots altogether. However, like Allingham pointed out, a robot with high-speed capability still may be your best bet. Even if you’re not running the robot at high speed now, it may come in handy for future improvement or redeployment to another task.