As seen in SMT Magazine, June 2001 by Dr. George T. Ayoub, President & CEO, MVP Inc
While automated optical inspection has been integrated successfully into many production lines, much additional useful data for the inspection process remains to be used. How? Enter Statistical process control . . . and a “catalyst.”
Simply stated, statistical process control (SPC) is a method of monitoring, controlling and improving a process through statistical analysis. Its four basic steps include measuring the process, eliminating its variances to make it consistent and monitoring and improving the process to its best target value.
Throughout electronics manufacturing, SPC has been implemented with overall success. Quality generally has improved while the level of accountability has risen as data (some of which having been collected by automated optical inspection [AOI] systems) are on hand to monitor defects and to indicate less than optimum system performance. However, an inherent flaw in the general use of SPC exists: the data are not real time. Typically, the information is analyzed, line problems are discovered and changes made – a process that is hours and sometimes days after the fact. During the time between identifying the fault and correcting it, thousands of imperfect boards may have been produced at high cost and rising scrap levels.
In a world of tight profit margins where getting it right the first time is critical to the OEM’s or contract manufacturer’s bottom line, it simply is not acceptable to waste time and resources manufacturing faulty printed circuit boards (PCBs) or operating sophisticated SMT lines inefficiently. Thus, in a perfect world, SPC would be real time. But how would it work? Ideally, when a process deviates from a preset specification, the operator would immediately be notified and a plan of action suggested for instant correction. Above all, the fault would not be identified the following morning at the daily SPC meeting. The key to such results is the coupling of sophisticated AOI and SPC methodology: using traditional inspection information as a key tool on the production line. The goal of real-time AOI/SPC is higher throughput and yield, greater line efficiency, and lower overall manufacturing cost.
♦Linking AOI and SPC
Using the standard defect data and variable measurements that today’s sophisticated AOI systems can generate, it is possible to create a real-time inspection/defect system that alerts line operators whenever any process exceeds preset limits – and does so immediately. This, in essence, is real-time SPC: a system that continually monitors line performance, detects problems with each board, checks pick-and-place operations, measures feeder and nozzle functioning, and stringently controls process variability to optimize performance, throughput, first-pass yields and overall quality.
Two factors are key to real-time SPC:
- Fast, accurate AOI, which can take pre- and post-reflow solder measurements
- A series of “simple” line controllers that use an RS485 network to collect data
The latter are immediately sent to an intelligent information maintenance network that translates the circuit references into specific machine, feeder and nozzle information. The network also permits the transfer of data from the screen printer, the pick-and-place machine, the reflow oven and the AOI system.
The result of such integration is improved line performance and the beginning of an expert response system that can identify problems created by individual handling equipment. Operator instructions and corrective actions directed to the “guilty” machine immediately follow, based on process-control limits set by engineering. Other essential data, e.g., best and average cycle times; run, blocked, “starved” and down times; and the top 10 feeder and nozzle problems can be captured and analyzed in real time.
The result is termed Dynamic Process Control* (DPC) because, in operation, it goes beyond real-time SPC by adding audible alerts and, via a visual display panel, proactively suggests what corrective actions should be taken. Generally, there is no need for control charts and procedures because the sample interface on the machine is displayed uniformly on all machines. Additionally, DPC may be helpful in improving quality control, real-time line balancing, inventory control and production forecasting.
♦What DPC Can Accomplish
By calculating best and average cycle times, manufacturers can know, often for the first time, just how much system use actually is being achieved. It is common for operators to believe that their lines are reaching upwards of 75 percent utilization, yet have no real-time data to substantiate this belief. Thus, in some cases, the addition of a DPC system might reveal that use often has been much less than 75 percent; but, at the same time, immediately find faults that suggest corrective action and quickly helps return lines to full utilization.
For example, DPC can be used to improve pick-and-place equipment utilization by tracking faulty nozzles and feeders back to the manufacturer, or faulty components to a specific reel, thereby quickly optimizing a line to meet a changing product mix.
Extra modules have been developed to perform other real-time tasks, such as setup verification to eliminate misleads by validating acceptable parts numbers for each reel load. Other such modules include works in progress (WIP) tracking to log the progress of panels through the process; materials traceability to trace field faults back to a particular component from a specific manufacturer; materials management to track material consumption and warn of impending shortages; and inspection feedback – a loop between AOI systems and placement machines that alerts operators to potential feeder problems.
Finally, DPC can be used to verify line setup to build products right the first time, avoiding expensive repairs and increased operational costs. Because a standard database stores information, engineers and manufacturing personnel easily can run simple reports using standard tools such as Access. Data also can be integrated into a customer’s own standard information collection and reporting system.
♦What AOI and SPC Lack
Combining AOI data with software and data collection modules placed throughout an SMT production line can turn static SPC into real-time DPC. In operation, all system warnings are driven by preset limits. Specific corrective actions are recommended with instructions flashed on the data collection modules. Systems can be referred down to specific machines, nozzles and feeders from an online computer anywhere in the world.
Although DPC can function as a real-time “report card” by showing how well lines and machines are running and delivering the data to prove it, it is not about finding fault. Rather, combining AOI with hardware and software to gather and analyze data becomes another tool in the quest for optimum quality and the best possible line performance.
Given the need to speed time-to-market and the economic necessity of building products right the first time, standard AOI and SPC can no longer provide the full range of information (downtime, uptime, delay time, etc.) needed to optimize line performance and to guarantee quality levels, much less do so in real time. Not long ago, before the almost universal adoption of AOI and SPC by electronics manufacturers, quality levels and line optimization were merely “guesses” – those based on “anecdotal” evidence and after-the-fact field failures. With AOI, defect detection became much more precise. Manufacturers could, for the first time, “see” the problem in real time. And with SPC methods, line optimization and quality control were improved. For the first time, engineering had hard data that could be used to monitor the line and adjust it for optimal performance (albeit after the fact).
For its part, DPC promises to complete the picture by changing the face of traditional AOI and SPC.
SMT: Dynamic Process Control is a trademark of Machine Vision Products Inc.