The Pick and Place Machine Unveiled
When the printed wiring or circuit board was first created in the 1940s, it was done so in part to reduce the increasing complexity of attempting to wire circuits manually. The PCB fulfilled its role perfectly and was an overwhelming success, since becoming a multi-billion dollar industry. However, as PCBs became more prevalent and advanced, it also became difficult to quickly and accurately produce populated circuit boards in large quantities. This need to populate the tens to hundreds of thousands of circuit boards lead to the advent of pick and place machines, an automated method of placing a large quantity of components on boards extremely quickly. While simple in concept, the pick and place is an amazingly complex and precise machine. The variety of situations in which the pick and place is used has created an interesting smorgasbord of styles for the pick and place machine, though all operate according to a set of basic principles are generally unchanging.
Automated pick and place machines all function by picking up a component in one place and putting it in another. Instead of using fingers or tweezers like a human operator would use, the pick and place uses a small suction nozzle that presses against the component, using suction to hold onto it. Once the nozzle has a firm hold of the component, it lifts it up and places it where it is needed, releasing the suction as soon as the part is in place. While the original and final position of the component are programmed into the pick and place machine, or PnP, minor variations in location can occur. Then, these locations are confirmed with computer vision, verifying that the board is truly in the expected position. It also verifies that the component is actually picked up by the nozzle. Fiducials and tooling holes are mentioned frequently in discussion of automated machines, as both are used as references by the computer vision. Fiducials are created solely for this reason while tooling holes are created primarily to mechanically hold boards or panels in position though they can also be used as reference points.
The vast majority of pick and place machines are based on a Cartesian coordinate system, with most motion available to the head moving in the X and Y directions. A limited range on the Z axis allows for the nozzle to go up and down to select components, however, most movement is two dimensional. Particularly in the low-cost, do-it-yourself community, there is currently experimentation on creating delta PnPs based off of similar designs for 3D printers. These are gaining popularity, as they are mechanically simple and low cost, though the mathematics behind the movement is marginally more difficult. They have certain limitations, such as limited range of motion. As the X and Y-axis requirements increase, the delta PnPs start to become significantly taller to accommodate the side-to-side motion, which makes the entire machine somewhat unwieldy and the arms become longer, heavier, and more prone to errors. Also, as the delta motion is a lower cost machine, it is less likely to have a rotational feature on the tip to align components as needed as well as other features. The Cartesian based systems are much easier to scale and the two dimensional nature also aids in quicker movement. Also, as it is the older, more established technology, it has had more time to be optimized for speed, accuracy, and cost.
Pick and places are not limited to a single nozzle. The lowest end pick and places will still likely offer different nozzles that can be attached to the head, though the swap will have to be done manually. However, for higher end machines, the pick and place can use multiple heads at a single time. This can be set up differently, such as by switching nozzles to handle a larger or smaller component, or grabbing more components of the same size at the same size by using different nozzles. There can be quite a bit of flexibility in the choices. With this flexibility also comes a certain degree of challenge in selecting and utilizing the nozzles correctly. Their size needs to be taken into account in the setup of the system, they must be small enough to attach to the smallest parts but large enough that they can hold the weight of the larger parts. They also must be well grounded to prevent small parts from becoming electrostatically connected to the nozzle instead of being left in their appropriate place on the board. Solder paste is helpful in these situations as it acts as a glue to hold the component in place.
Solder paste can be placed on the board in one of a variety of ways. The most common for large production runs is the automated usage of stencils. Stencils are made by cutting holes into either a metal or plastic sheet in the places where solder is needed on the board. These are created with the information in the solder paste gerber file. Before entering the pick and place machine, this stencil is placed on top of the board or panel of boards and properly aligned so that the holes in the stencil match up with the pads of the board. A large dollop of solder paste is placed on the top of the stencil which is then dragged across the stencil with a squeegee. The solder paste goes into the holes of the stencil, covering all of the pads. However, for very small prototyping runs when a stencil is not worth the time or money, some PnPs are equipped with solder paste depositing nozzles that will deposit individual dollops of solder paste on each pad. This is slower on a per unit scale but saves on cost and has a shorter waiting period for the arrival of the stencil after ordering.
In a full production line with the correct board design, bare boards can be placed in one machine and they will come out the end fully populated and soldered. They can be pasted, populated, and soldered without any human intervention. This movement through the machines and from one machine to the next is done with a rather straightforward conveyor system. The boards are moved into place, processed, and then moved automatically to the next step. Frequently these same machines with the conveyors will not actually be connected so that while a human operator feeds boards into and out of each machine, they need to be manually moved from each stage in the process to the next. The smaller the system setup, the more likely that each machine, including the PnP, will be a batch process versus conveyor. In these cases, the boards are manually place in the exact position where they are needed, they are populated, and then the boards are removed to be replaced with the next board.
Another feature of pick and place machines is how flexible they are in their source of components to place. Anyone who has placed an order with any electronics distributor has noticed a myriad of different ways to receive their parts, predominantly Tape and Reel and Tube, though there are variations on these such as cut tape or custom reels. Tape and Reel as well as Tube are very popular with distributors because they are ideal for usage in PnPs. Reels of parts have small holes on one side of the tape, much like one would find with the paper used in very old printers. These holes are used by the PnP to move the tape forward as parts are pulled out and the tape needs to be advanced. The clear plastic is also pulled off the tape to reveal the parts before they’re picked up by the PnP. This plastic removal is done either by a motor with tension feedback or manually. A common trick when working with less expensive machines is to actually tie fishing weights to the plastic to use gravity to pull off the tape. Tubes are simpler in their usage but non-ideal for smaller components. Tubes are placed in the machine slanting downward toward the PnP to let gravity pull each component into place after the previous one is taken. So that the components don’t get stuck to the tube, a small vibrating machine is used to encourage the parts to slide down. Most pick and places are capable of pulling parts from other sources, including trays or even custom plates. This is done instead of requiring more manual oversight of switching out the trays. For small runs, this is not as much of a concern.
Pick and places are impressive tools but they still have their limitations and drawbacks. As with any computer system that interfaces with physical reality, there are sometimes miscommunications. Occasionally, pick and places do not pick up their intended part, or lose it in transfer, or do not drop off the part at the intended location. There are many steps to reduce the likelihood of this happening, such as human supervision and computer vision. However, these precautions are meant to correct mistakes instead of prevent them. Also, with many parts, in particular the small passives, once they are dropped, they are essentially gone. Their size and cost do not warrant a search and replacement and they are typically discarded. It is due to this that assembly houses request overages in parts that are provided. If the PnP is not fed all of the necessary information, it may also have the issues of hitting objects that protrude vertically from the board. This could cause serious damage and misalignment to the machine and, while not common, needs to be actively avoided.
Because of the large variety of options in pick and place machines, many designers and engineers are uncertain of what to do in their situations. The first question is whether or not to get a pick and place machine, to do the work by hand, or to get the boards assembled at a PCB assembly house. This is highly dependent on the time, budget, and experience constraints. For high school or college students on limited budgets, hand population and soldering may be the best option. The difficult decisions come where there’s a bit more money and a little less time.
If equipment is priced reasonably, it is easy to think that it may be worth it to purchase and operate a PnP internally versus having the assembly work done by others. This is a valid option and in certain circumstances, may be entirely warranted. However, many small businesses fail to take the lifetime costs of owning a PnP place machine, which includes far more than the machine itself. The infrastructure that may be built to support it, the training required to adequately operate it, the maintenance and repair of the machines, certification and waste costs, and the personnel costs of operating it all come into account.
The benefit of owning a PnP machine in-house is increased flexibility, decreased time explaining a project to someone, and a potentially lower lifetime cost. If the PnP is merely for prototyping or <100 piece production runs, then a smaller, batch PnP with a single head will likely be sufficient to cover all your needs. More emphasis should be placed on the ease of changing out a design and flexibility on how to accept small quantities of parts. If this PnP is part of a new and large manufacturing process for large quantities of only a few types of products, then there will be more focus on outright speed optimization. In that case, it is recommended that you hire consultants to help with setup and in finding the right personnel to handle the manufacturing.
Pick and place machines have been an instrumental force in fueling the electronics revolution, enabling designers and manufacturers to be able to produce large quantities of electronics in very short amounts of time. Identifying their strengths and weaknesses, knowing what to use and when, and understanding the respective impact on all facets of design and logistics will allow for better usage of time and resources. Properly balancing the different parts of production and prototyping will decrease bottle-necking and make reaching deadlines a more achievable goal.