Quality Test & Inspection: Sensor-Driven Error Proofing
When it comes to the level of quality that is acceptable in today’s manufacturing marketplace, one fact stands out. State-of-the-art customers are not accepting less than state-of-the-art products from their suppliers. Advanced manufacturing operations can only meet the requirements of these customers by making 100% perfect parts and products at constantly reduced cost and productivity.
Sensor-Driven Error-ProofingDuring the past few years, the terms total quality management, lean manufacturing, error proofing, mistake proofing and Poka Yoke have been used to describe various programs designed to drive out mistakes in the manufacturing process.
Today, it is becoming recognized that the most effective error-proofing programs use sensors to eliminate errors as an integral part of the manufacturing process itself. Old-fashioned quality checks, particularly after parts or products are completed, no longer cut it. Whether the production process is automatic or semi-automatic, using sensors to verify the manufacturing steps that drive quality work is the best way to guarantee that the part being produced meets customer specifications.
• Discrete sensors. A wide range of sensing technologies is used in automated assembly processes. Comparatively simple discrete, or on-off, devices such as inductive proximity sensors can be used to verify parts presence, making sure a part is in the correct place at the right time before the next manufacturing step occurs. These sensors also can detect or confirm features such as tabs, nuts and flanges, as well as validate hole presence or absence and correct part nesting. Tubular, flatpack and block-style sensors in a wide range of geometric configurations can be used to meet specific mechanical or electrical application requirements.
Discrete photoelectric sensors in thru-beam (energized pairs), retro-reflective (used with a dependable target, such as a reflector) and diffuse reflective (self-contained emitter-receiver pair) configurations are all used to detect parts presence at more than 30 to 40 millimeters or when nonmetallic components must be detected. Inductive sensors require a metal target.
• Analog sensors. Analog sensors provide continuous voltage (0-10V) or current (4-20mA) output to the control, allowing precise gaging, measuring, and positioning of parts and processes.
When a few millimeters of sensing range is required, use either analog inductive proximity sensors with metal targets, photoelectric sensors with either laser, red or infrared emission, up to 350 to 400 millimeters of sensing range.
Long-range inductive proximity sensors-up to 50 millimeters of range for metal targets-and long-range laser sensors and Time of Flight Laser (TOF) sensors offer up to 6 meters of continuous voltage or current feedback. Several models of both inductive and laser technologies offer discrete on/off set points that can be programmed anywhere within the range of the sensor. This feature can help to establish the next production step go/no-go parameters, as well as correct position feedback all by a single sensor device.
• Beam-break sensors. By combining the emitter and receiver of separate thru-beam sensor pairs, it became possible to create a new sensor style in both U-shaped and L-shaped versions more conducive to common error-proofing applications such as part presence, height verification applications and on feeder bowls for component flow validation. This new style of sensor eliminates the drawbacks of thru-beam fiber optics commonly found in assembly and automation operations-pairs going out of alignment, cumbersome bend radii issues and vulnerability to mechanical damage.
Color sensors are available in a range of sophisticated, relatively high-cost models, but there are low-cost devices available that will detect and hold three colors or shades of color in memory and that are easy to program and cost-effective. These solve 80% to 90% of the day-to-day color sensing requirements found in manufacturing processes.
• UV tracing. UV sensors, also known as luminescence detectors, are unique animals. These devices are 100% reliable because they only respond to the ultraviolet spectrum of light. However, it is necessary to make sure that dependable target material is applied to the part needing to be sensed. Target marking materials can be water- or solvent-based and are sold in powders, inks, markers and crayon-like formats. The benefit is that they are invisible to the human eye, inert and have no negative impact on product aesthetics. Moreover, many greases and lubricants inherently glow when exposed to UV sensors. In fact, many engine test stand and powertrain manufacturers use UV sensors to detect leaks and to determine fill levels in lubrication tests prior to installation.
Tracking Build InformationFlexible manufacturing requires the ability to construct various product variants on the same line. Because different product versions have unique features to error proof, the exact version being manufactured must be identified and traded. This is accomplished most effectively by radio frequency identification (RFID) systems that store and build data on a small data carrier, sometimes called a tag or chip, affixed permanently to a build pallet.
Before assembly begins, the data carrier is loaded with the build information that will instruct all downstream processes as to the exact version being manufactured. Correct assembly is verified by comparing the build information to what the error-proofing sensors detect.
ID systems should be matched to the level of complexity required for the application at hand. For simpler assembly operations, a read-only, 8-bit parallel system works fine. In this case, all that is needed is a self-contained read head and matching carriers connected directly to inputs on the control system. Information is referenced centrally.
When build information requires a decentralized approach, read-write systems are appropriate. Before assembly begins, the build information is written to the data carrier. The assembly system reads the build information at each station to determine what assembly and error-proofing operation is required. In addition, the actual test results can be loaded into the data carrier for subsequent archiving. These systems read and write data through standard interfaces such as ProfiBus, Ethernet and DeviceNet.
Metal-Forming OperationsThe same error-proofing principles hold true for metal forming as they do for assembly and automation practices, although there are a few additional objectives. In conjunction with stamping perfect parts with zero bad parts per million, additional goals are to protect dies from die crash damage, prevent die lock up and ensure production without interruption.
Discrete sensors, both inductive proximity and photoelectric, are used to monitor stripper position, strip feed, pilot-hole detection and feature detection. They also can be used in dies for slug-out and parts-out sensing.
Short-range analog sensors, inductive and photoelectric, are used to measure bend angles on parts, check for specific features and measure other critical dimensions. They are also used for nut detection on stand-alone, error-proofing stations or where value-add-such as welding or assembly on-die-takes place during the stamping process. Analog inductive sensors often are used to measure press parallelism. Photoelectric sensors can be integrated to precisely measure roll feed at long range, plus parts out and slug-out detection.
Rapid die change can be accomplished with power remote sensors. Here, two specialized inductive components pass power and signal to up to 2 amps and up to 64 sensors on a single run. The information and power is passed through an air gap and without a mechanical connection, useful where any machine sections must be in component pieces but act as one. This remote sensing system prevents damage to dies and operator fingers during die change and shortens the time to swap out dies.
RFID sensor systems also can be used for die identification and storage tracking, crucial to customers with many constantly used tools in house. RFID in conjunction with tagged dies can ensure all die segments are correctly in place in the press to protect against die crashes before the stamping process occurs.
Welding CellsOften, robotic or semi-automatic welding cells operators are overwhelmed by excessive consumption of nesting, clamp and unclamp indication sensors, error proofing and feature-detection sensors because of the inherent high temperatures and the constant attack of weld slag on conventional sensing devices, particularly if the devices are not protected properly or if the sensors installed were not designed for weld environments.
If all parts to be metal inert gas, Tungsten inert gas or resistance welded are not joined correctly, the potential for scrap, or worse, mal- produced parts ending up in a customer’s hands can be the result. Rework and sorting can be costly. Misapplied sensors can cost the average weld cell operator anywhere between $6,000 and $64,000 per month, not counting maintenance downtime.
As in all manufacturing processes, it is critical to match the sensor to the application. To keep weld cells functioning correctly, inductive proximity nesting validation sensors must be able to withstand the rigors of loading impact. Coated inductive proximity types must be able to keep elevated temperatures at bay while resisting accumulation of weld debris. Photoelectric sensors must have the high excess gain necessary to sense through smoke and oily film. Clamping sensors must be able to handle all of the hostilities listed above, plus shock and vibration.
Whether an industry discipline revolves around assembly and automation, metal forming, or automatic, robotic or semi-automatic welding, the need for error-proofing the process is imperative. There are scores of cost-effective, off-the-shelf sensing choices for integration into processes that can achieve 100% inspection of parts, detecting parts presence and feature detection, measuring critical dimensions and following products through entire production processes with high reliability. All of this results in a more satisfied customer base, a less stressed workforce and increased profitability for any manufacturing organization. Q
Dave Bird is business development manager at Balluff Inc. (Florence, KY). For more information, call (800) 543-8390, e-mail email@example.com or visit www.balluff.com.
Tech Tips•Many error-proofing programs use sensors to eliminate errors as an integral part of the manufacturing process.
•Matching the sensor to the application helps ensure quality.
•Misapplied sensors can cost the average weld cell operator anywhere between $6,000 and $64,000 per month, not counting maintenance downtime.