Edited by Leslie Gordon
Our company makes video cameras that monitor hazardous enclosures such as on nuclear reactors and off-shore oil rigs and we continually develop equipment to automate our in-house assembly, forming, machining, powder coating, and injection-molding processes. The equipment performs secondary operations, robotic welding, and other specialized tasks. We also design and build automation systems including a chemical-process system for a metal finisher and equipment for waste treatment. Time has taught us that automation doesn’t just come from pressing a green button — automation must be a mind-set.
A case in point comes from an early purchase of a robotic-welding cell. We designed the fixturing with help from the robot manufacturer by sending .STP files of tooling based on how the parts to be welded were positioned and clamped. This data also helped in determining cycle times and fixture population. Everything looked good, so we fabricated 10 fixtures and shipped them to the cell manufacturer, who coded programs and made some acceptable parts. We figured that once the robot cell hit the floor at our facility, we could move directly into production.
After the cell started making parts, however, the robotic fixtures gradually fell apart because of crashes and other operator-related issues. Weld programs somehow got altered and deleted. Management did little to investigate or question the deteriorating project. For the design group this was disheartening, to say the least.
The situation was perplexing. From the beginning, management seemingly accepted and sponsored the project. But now, they didn’t seem to know how or even want to make it work. The painful lesson: When people in charge are not willing to manage automation on an ongoing basis, more than likely the project will fail. The welding system is currently idle, but it does make a fine coat rack.
The upside is we gained valuable insight and now ensure all stakeholders are on the same page before proceeding. Everyone has a defined role with specific responsibilities and is held accountable. With this learning experience under our belt, we are rethinking the welding application and looking forward to its eventual success.
The right way to automate
Compare this dud of a job with a recent project that everyone from management to cell operators fully supports. We manufacture a camera enclosure from aluminum extrusions and die-cast end caps in a variety of length and body configurations. Originally, workers manually applied sealant to the end caps and assembled units by hand. But when demand for the product increased from about 450/week to 450/day, production suddenly took six to eight people working 10-hr shifts. At these volumes, manual assembly was inconsistent and units needed lots of hand touch-up and quality inspection.
Consequently, everyone understood the need to automate. The design team purchased a small, off-the-shelf Scara robot, stepper-motor- driven dispenser, and motorized conveyor. And we designed an air-driven rotary actuator with vacuum fixtures to place and hold the die-cast parts, and an assembly station. PLCs control the rotary fixture and assembly station and communicate via the Ethernet. We can also monitor the entire work center remotely via the Internet. A large flat-panel monitor displays data such as parts/hour and parts/day, giving management and cell operators constant, real-time production information. The automated cell now produces up to 100 units/hr with only three operators.
Our company fills its automation needs with just two members of the maintenance department. These individuals come up with ideas for equipment and fixtures, build the CAD models, spec and purchase materials, machine parts, as well as wire, assemble, and program. Fortunately, a well-equipped machine shop and sheet-metal department lets this all happen without outside contracting. The capabilities also dramatically reduce ROI time.
In addition, in-house CNC equipment and programmers play a major role in getting projects approved. The capability to generate .DXF and .IGS files from CAD and have robotic or NC programs written the same day is a distinct advantage in implementing automation.
CAD also comes in handy for motion analysis, mass calculations, and 3D simulation to verify design intent. Additionally, many suppliers provide sites or software with 2D and 3D component models. Many of these can be dragged and dropped into a design during a virtual build.
When designing automated systems, a significant challenge is getting all the information needed to best understand requirements and anticipate results. In large companies, it’s sometimes difficult to obtain information without a committee, but accurate data helps ensure a successful project.
Another challenge is to keep things simple and resist over-engineering. A good design is one that’s reliable, straightforward to use, and easy to repair and maintain. Whenever possible, we design equipment with as much flexibility as possible to allow for future or alternative applications.
Also difficult: justifying the additional expense of aesthetics, especially for one-off machines. But building a visually pleasing machine can actually help keep it running. Operators are more inclined to take care of equipment and report problems when a machine looks like it was purchased from an equipment manufacturer, not just cobbled together.
Because we make custom automation systems throughout the world, another challenge is complying with RoHS and Reach standards. It is necessary to consider energy conservation, and incorporate auto power-down or sleep modes, state-of-the-art energy-efficient motors, and other power-conservation technologies into equipment designs.
Of course, every project has time, costs, and budgetary constraints. The best approach is to create standard practices and uniform design guidelines. For example, in programming, we write a standard for E-stop circuits. From then on, we use the same design over and over. We don’t have to decide what an E-stop means. We also try to use standard-size material.
Also, it’s a good idea to use the same component manufacturer and common parts whenever possible. Creating a library of common components makes it easier for a designer to decide, for instance, to use a 1-in. bore, 2-in. stroke cylinder — when the application might call for a 1.75-in. stroke — because this common component can be used on other pieces of existing equipment that use a 2-in. stroke cylinder.
Overall, building and implementing in-house automation takes a lot of time and resources. Start with a flowchart or box diagram if the new design is to replace an existing system. Thinking ahead, take before and after photographs of existing lines or layouts. Make sure to get management and operator input so everyone buys into the design or concept. Keep notes on the design and design intent.
Also, make sure that every machine has a least two manuals — one for operators and one for maintenance that includes all electrical diagrams, ladder logic, BOMs, part drawings, and the most recent backup of the processor ladder and OIT programs on any install, no matter the size or complexity. We always program diagnostics into the controls to aid troubleshooting. Troubleshooting data is password protected with several levels of security to allow the proper access for trade-level competence.
In addition, every new piece of in-house automation equipment connects to our local network. At times, the design team camps out at a new install for a week, working with supervisors and operators until everyone is comfortable and well-schooled. As designers, what we assume is common knowledge is not necessarily easily understood by everyone. Here, it is helpful to randomly select staff to review documentation and procedures. This practice provides information that helps us successfully implement and sustain automation.