It's been said that, with the rise of modern telecommunications, we're living in an ever-smaller world. True or not, it's undeniable that plastic parts keep getting smaller, as their use in the medical and electronic industries continue to...
It’s been said that, with the rise of modern telecommunications, we’re living in an ever-smaller world. True or not, it’s undeniable that plastic parts keep getting smaller, as their use in the medical and electronic industries continue to grow.
The term “small part” can mean different things to different people, however. Engineers define “small” as any component less than one inch in length. For plastics processors, a small part – or miniature or micro part – is much smaller: No bigger than 0.25 inches in overall size, and sometimes a lot less than that. (Think grains of sand.)
Straight away, it’s not hard to spot the processor’s conundrum: How do you handle a part that’s barely visible to the human eye?
There are several options, beginning with the most basic and inexpensive – the human hand. The downsides? Clumsiness and slow, inconsistent results. And pity the poor shop floor operator. “The pressures of finding and handling very small parts without damaging them can make an operator’s hands literally shake,” said Mike Cicco, general manager of material handling for Fanuc Robotics America Inc. “It’s definitely not something you want to do for an eight hour shift.” Still, the appeal of the hands-on, zero-investment approach is strong – in the beginning, anyway. “Virtually every processor that decides to invest in automation for small parts tried first to do it by hand, and simply found it too difficult and time consuming,” Cicco continued.
For processors ready to move beyond the tongue-moistened finger approach, choosing an automation method for small parts requires some thought about balancing speed, cost and quality, as well awareness of the shape and size of the parts to be handled.
A first approach is to custom design and build a system specific to the process at hand, complete with dedicated end-of-arm-tooling – called a “hard” automation arrangement.
A second option is a robotic automation system engineered for microassembly. “Robotic automation gives the processor a level of flexibility that ‘hard’ automation doesn’t,” Cicco said. “As the requirements change, the system is changed too, sometimes with as little as a ten-minute program change.” In this approach, the robot is a flexible tool, complete with a programmable controller. After the program is written, the end user or a systems integrator can then develop end-of-arm-tooling (EOAT) to grip the item to be moved. The good news, from a programming point of view? The size of the parts being handled usually doesn’t affect the programming process. “The robot has its assigned field of view, and within that field of view the only thing that matters are pixels. By that measure, even the smallest part can be made to appear large,” Cicco continued.
Portable side-entry robots offer another alternative to the “hard” option. “Yushin’s SXA-10II side entry robot interlocks with super-small molding machines with which product take-out and installation of a take-out robot have been considered difficult, and allows you to raise the ratio of good products and achieve stable small part molding,” said Brad Lemieux, sales manager with Yushin representative En-Plas Inc.
Whichever option the processor chooses, a key to a good robotic automation solution is a vision-identification system, wherein the camera looks for parts of a specific shape, size or color. The user defines a primary pick point in the workspace and the camera locates the part. The robot then moves in and – gently! – picks the part up. “We’ve designed many EOAT and downstream solutions for small part applications that use a vision-identification system for part verification or other means to prevent rejects getting mixed with good parts,” Lemieux continued. “We’ve even had a two-shot machine application where our customer had extremely stringent part deviations they had to be within, and we were able to offer them a turnkey solution that implemented a scale that could decipher 0.001 gram difference from shot to shot.”
And whether it’s a “hard” solution or not, there are a slew of EOAT components your system might incorporate – including grippers, vacuum cups, and bowl feeders – as well as sexier equipment like industrial robots.
A wide range of specialized grippers are available to handle small parts, beginning with mini-finger grippers featuring jaws that can be machined down to very tiny shapes to fit the part, whether it be an interior or exterior grip.
Grippers offer greater control of forces than conventional tweezers. “A big benefit of a gripper is that it can accurately locate the part,” said Dr. LaRoux Gillespie, a consultant and author of numerous books on automation and precision machining. “Once the gripper has the part between its fingers, the processor knows exactly where that part is.”
Be careful, though: With a pneumatic gripper, it’s possible to produce enough force to damage the part. Force sensors can be used to signal the robot that it has applied enough force to hold the part. Stroke position sensors are helpful in providing feedback to the robot so it knows whether the gripper is open, closed or in transit, Gillespie continued.
Also, since small parts are often presented in arrays spaced closely together, the grippers must be very small themselves to work effectively in these tight matrices. “Gripping and releasing small parts is typically more challenging because of the accuracy, repeatability and speed involved,” Gillespie said. “As a general rule, higher quantities require lower cycle times.”
If there’s a risk of part damage due to the force applied by a gripper, or if precision is not a big concern, vacuum probes are a good option for small parts handling. “Vacuum probes find frequent application in the mass production of flat parts and for parts smaller than wooden match heads,” said Gillespie. “It’s a method that processors like to use because it’s cheap, clean, fast, and readily reconfigurable.”
The drawback to vacuum probes? “Vacuum systems are often unable to handle certain small part shapes unless the tips are modified,” Gillespie continued. “And their usefulness is limited when parts must be pushed on or pulled off fixtures.”
A third option involves vibratory bowl feeders. The vibration of the bowl automatically orients the parts, which are then transported via rails to the next station. The rails and gates used can also perform simple inspection tasks, such as checking that external features are present and automatically rejecting nonconforming parts. Bowl feeders are standard, off-the-shelf devices, Gillespie said, but the systems’ orienting and part-rejection features must be customized for a specific part or part family’s shape and weight. “Vibratory bowl feeders and their associated rails are one of the most cost-effective approaches for processing large lots, ranging from 3,000 to one million parts or more,” he noted.
Drawbacks? “For smaller lots, the requisite setup time normally outweighs any savings in handling time,” Gillespie said. “Contamination from application to application is also possible, since part fragments might become embedded by the vibrations and rub into any subsequent production runs.”
“In my mind, robots are useful for small parts handling when a processor will be making hundreds of thousands of parts, in production runs that will last for years,” Gillespie said.
Since the normal robotic rules don’t always hold in the microworld, Gillespie and others stress that processors have to understand upfront what’s important – and what’s not. “In small parts robotic automation, payload is not an issue – reach is far more important,” said Mike Cicco. “A processor should tailor their choice to the s
mallest robot with adequate reach, because the smaller the robot, the faster and more repeatable it is.”
In the end, the best configuration to use depends largely on the application and the work envelope, Cicco continued, with other considerations including process parameters, part weight, and repeatability.
The good news: Thanks to recent advances in hardware and software, insiders say, nearly any style of robot – parallel link, articulated arm or SCARA – can be used in small parts handling.
Parallel link units – in which the end effector, or tool, is linked to the power source via parallel links, or beams – tend to be found in standard assembly and food processing applications, and are making the leap to small parts handling. Fanuc’s M-1iA lightweight six-axis, parallel-link robot can be installed in a variety of orientations, and has a three-axis articulated wrist to enable a 3-D part movement. The M-1iA also offers a four-axis model for simpler operations such as part picking or kitting, and a camera for iRVision can be integrated into the robot. “This system can do very complex, small things in a very small, 10-inch diameter work envelope,” said Cicco
In articulated arm models, the robot arm has at least three rotary links and often rotates on a base. According to Cicco, an articulated arm robot such as Fanuc’s LR Mate 200iC can provide the same repeatability as a parallel robot, carry a larger payload, and offer a larger work envelope – albeit at slightly slower speed. “In general, an articulated arm robot tends to be more flexible than other robotic styles, allowing it to better accommodate job changes,” he said.
SCARA robots have two parallel rotary joints that determine the X and Y positioning, as well as a sliding shaft at the end of the arm that moves in the Z-axis. For high speed, pick-and-place work with small parts, SCARA units are said to really shine. “SCARA robots are the most precise, cleanest and fastest robots for small-part assembly work,” said Gillespie.
Downsides? Less flexibility than an articulated-arm model. “If a processor only wants to pick a small part up and move it on the X and Y plane, SCARA robots are a very good option – they’re cheap, fast and accurate,” said Mike Cicco. “In my experience, however, a processor who wants three axes of motion will find it more cost-effective to purchase a pre-designed engineered robot, because of the amount of additional programming and extra engineering work required to get a pick-and-place robot to that performance level.”