Early development (1960s–1970s) The origins of pick-and-place automation can be traced back to the 1960s, with the introduction of the first
industrial robots designed to perform repetitive tasks with high precision. The
Unimate, developed in 1961 by
George Devol and
Joseph Engelberger, was the first programmable pick-and-place robot used in manufacturing environments. Early applications included
material handling,
assembly, and
packaging, laying the groundwork for later advances in
surface-mount technology (SMT) and
electronics manufacturing. Subsequent innovations, such as the
Stanford arm (1969) and the PUMA (robot) (1978), further expanded the capabilities of pick-and-place systems, enabling greater flexibility and accuracy in automated production environments.
1980s and 1990s During this time, a typical
SMT assembly line employed two different types of pick-and-place (P&P) machines arranged in sequence. The unpopulated board was fed into a rapid placement machine. These machines, sometimes called
chip shooters, place mainly low-precision, simple package components such as
resistors and
capacitors. These high-speed P&P machines were built around a single turret design capable of mounting up to two dozen stations. As the turret spins, the stations passing the back of the machine pick up parts from tape feeders mounted on a moving carriage. As the station proceeds around the turret, it passes an optical station that calculates the angle at which the part was picked up, allowing the machine to compensate for drift. Then, as the station reaches the front of the turret, the board is moved into the proper position, the
nozzle is spun to put the part in the proper angular orientation, and the part is placed on the board. Typical chip shooters can, under optimal conditions, place up to 53,000 parts per hour, or almost 15 parts per second. Because the PCB is moved rather than the turret, only lightweight parts that will not be shaken loose by the violent motion of the PCB can be placed this way. From the high speed machine, the board transits to a precision placement machine. These pick-and-place machines often use high resolution verification cameras and fine adjustment systems via high precision linear encoders on each axis to place parts more accurately than the high-speed machines. Furthermore, the precision placement machines are capable of handling larger or more irregularly shaped parts such as large package integrated circuits or packaged inductor coils and trimpots. Unlike the rapid placers, precision placers generally do not use turret mounted nozzles and instead rely on a gantry-supported moving head. These precision placers rely upon placement heads with relatively few pickup nozzles. The head sometimes has a laser identifier that scans a reflective marker on the PC board to orient the head to the board. Parts are picked up from tape feeders or trays, scanned by a camera (on some machines), and then placed in the proper position on the board. Some machines also center the parts on the head with two arms that close to center the part; the head then rotates 90 degrees and the arms close again to center the part once more. The margin of error for some components is less than half a millimeter (less than 0.02 inches).
2000s Due to the high cost of having two separate machines to place parts, the speed limitations of the chip shooters, and the inflexibility of the machines, electronic component machine manufacturers moved to all-in-one modular, multi-headed, and multi-gantry machines. These machines could have heads quickly swapped on different modules depending on the product being built, with multiple mini turrets capable of placing the whole spectrum of components with theoretical speeds of 136,000 components an hour. The fastest machines can have speeds of up to 200,000 CPH (components per hour).
2010 onwards Placement machines have developed all-in-one heads that can place components ranging from 0.4 mm × 0.2 mm to 50 mm × 40 mm. There is growing focus on software applications for the placement process. With new applications like POP and wafer placement on substrate, the industry has moved beyond conventional component placement. For many SMT users, high speed machines are not suitable due to cost. Recent changes in the economic climate have shifted requirements toward machines with greater versatility for short runs and fast changeover.
Developments until 2025 Since 2010, pick-and-place machines have undergone significant advancements driven by the demands of
electronics manufacturing services (EMS) and the proliferation of surface-mount devices (SMDs). Modern machines increasingly incorporate
machine vision systems,
artificial intelligence, and advanced
software for real-time optimization of the placement process. There is a growing emphasis on
flexible manufacturing and rapid changeover, allowing for efficient production of both high-mix, low-volume and high-volume assemblies. Furthermore, the integration of
Internet of Things (IoT) connectivity and data analytics has enabled predictive maintenance and improved overall equipment effectiveness (OEE) in contemporary pick-and-place operations. ==Operation==