Process Control

Estimated reading time: 9 minutes

STEP 2

Bob Kelley

John Weisgerber

Continuous quality improvement is a process control system that utilizes in-line inspection to help lower PCB assembly costs.

One of the fastest growing methods of process control is in-line vs. post assembly inspection. In-line inspection during SMT assembly is becoming a requirement to ensure the quality of electronic products and reduce the manufacturer`s scrap and rework costs. It is important to simultaneously meet the goals of producing a cost-competitive product while continuing to meet or exceed customers` quality expectations. An effective way to work toward this goal is to implement continuous quality improvement (CQI). Two of the basic principals in implementing CQI are to continuously improve all processes and to mobilize data and team knowledge to improve decision making.1 In-line inspection during SMT assembly provides the data and knowledge needed for this effort.

There are three major applications for inspection during SMT assembly: solder paste inspection before component placement; component placement inspection before reflow; and post-reflow inspection, which often includes some type of solder joint checking.

Inspection cycle time is critical for in-line applications. As manufacturing cycle times become shorter, it becomes more difficult to manually check each board for proper assembly. In-line automated systems must be fast enough to do the job within the allowed cycle time. Most people are not willing to increase overall manufacturing cycle time to do inspection.

Component Inspection Capability

When choosing a component inspection system, it is important to first define and understand process needs. The system should help to identify defects in the process and improve product quality. Begin with a survey of the current process. What are the most common types of component placement defects in the process? Next, be sure that the system has the imaging technology and features best suited to the company`s manufacturing process.

Inspection accuracy and repeatability must be suited to the process. Inspection accuracy for resistor and capacitor chips is usually not as critical as for leaded devices, and different systems offer different capabilities. Some systems attempt to identify defects based upon the use of a "golden board." In these systems, defects are identified by comparing differences between the "golden board" and the board that is being inspected, regardless of whether the differences are defects or are just harmless variations. In general, systems that provide real measurement data are better able to provide the defect detection and process control required for component placement. In addition, it`s easier to verify the performance of these machines to measurement standards.

A critical aspect of any inspection system is robustness. Circuits often have 1,000 or more components per board. If the inspection results in too many false calls, line operators will have to check the system results too often and will quickly lose confidence in the inspection results. Poor component placement inspection robustness is often caused by variations in the board or in the components. A system that is sensitive to changes in the color or finish of the printed circuit board (PCB), or that relies too heavily on special lighting techniques, may have problems when the board finish, lighting or color changes.

At a minimum, an inspection system should measure the position of each component in X, Y and theta, and must also check device polarity. Device positions should be compared to computer-aided design (CAD) data to see if the component position is within tolerances. Devices outside of tolerances should be identified and the measurements should be used to update statistical process control (SPC) charts.

Inspection Technology

There are two technology types available to perform in-line component placement inspection. Most systems use either gray-scale or color charge-coupled device (CCD) cameras. The cameras collect images of the PCB, and the images are analyzed to determine if there are any defects in each area. Camera-based systems can be very fast, but because they rely on the brightness of light reflected from the PCB, they can be sensitive to changes in lighting conditions or materials. Most systems that rely on cameras for image collection have programmable lighting to create optimal images of each site or component. As board complexity increases, problems with lighting contrast or shadowing may arise. As each image becomes more complex, the image processing becomes more difficult and the cycle times can drop.

Laser-based component inspection systems use a laser scanner to create a 3-D image of the PCB. This 3-D image is based on the height of the PCB surface and the components, and is much less sensitive to changes in component color. Laser scanning systems also can provide a 2-D gray-scale image, similar to the image from a CCD camera. This image can be used to identify objects where there is little height contrast such as board fiducials or identifying component leads in solder paste. Laser scanning provides accurate position measurements of the components. These measurements are important to help reduce the number of false calls and provide the information needed for optimal process control.

Cost Savings

There are several ways that an in-line component placement inspection system can pay for itself. Defects detected immediately after component placement are much less difficult and expensive to repair than the same defects found after solder reflow.

In-line process control is an important aspect of CQI and arguably offers the greatest potential for cost savings. Many available systems offer a variety of built-in SPC tools. It is important to correctly identify defects on each PCB, but process control capability can help eliminate the causes of component placement defects and reduce the number of defects actually produced. Few defects are created in the reflow oven if the solder paste deposits and component placements are within specification. A study by one manufacturer found that component placement accuracy faults accounted for 73 percent of their total component faults on 0402 packages.2 SPC methods can be used to trace problems directly to their source so preventative action can be taken. Real-time SPC charting can show the positions of components as they are placed to indicate when a system is drifting, or if the placements are becoming less precise. A consistent shift in one direction may indicate a problem with a motion system or a damaged placement mechanism. SPC tools such as Pareto charts can show the most common types of defects as well as identify the components on which they occur. Several defects occurring with a particular device may indicate a problem with a feeder.

Process characterization is another valuable service provided by component placement inspection systems. Testing the capability of a machine for its intended application is an important part of any new product introduction. Data from a proven, automated system ensures the accuracy and the amount of data required to make valid engineering decisions about equipment and processes. Some manufacturers use their automated inspection systems to verify the correct operation of each placement machine after it has been serviced.

Many SMT boards require high-speed, high-accuracy component placement. In-line inspection systems are needed to identify the defects as they occur and to reduce expensive rework after soldering. In-line inspection tools provide understanding of the placement process and the opportunity to improve SMT manufacturing yields. Detection of solder paste and component placement defects can help eliminate the need for expensive post-reflow inspection systems.

As a final note, regardless of the type of inspection equipment used, it is still only a tool. Inspection equipment should not be placed in-line with the expectation that process improvement will occur automatically and without effort. To achieve the highest quality as well as cost savings, a company must commit to the effort, staff and training that will allow them to use their data to build a knowledge base to improve production quality. Placing inspection equipment in-line provides important defect detection capability to eliminate costly scrap and rework. Even more importantly, in-line inspection equipment provides the right data and tools for SPC analysis and supports CQI.

REFERENCES

1 Michael Brassard and Diane Ritter, The Memory Jogger II: A Pocket Guide of Tools for Continuous Improvement and Effective Planning, Goal/QPC, www.goalqpc.com, ISBN: 1-879364-44-1.

2 Jesse Galloway, "Small Component Soldering Process Factors," Proceedings of the Technical Program, NEPCON West `99.

3 Don Revelino, "Achieving Single Digit DPMO in SMT Processes," Surface Mount International 1997.

4 Ray Prasad, "Pathways to SMT: Test/Inspection," SMT Magazine, October 1999.

BOB KELLEY and JOHN WEISGERBER may be contacted at GSI Lumonics, 4420 Varsity Drive, Ann Arbor, MI 48108; (734) 975-7600; Fax: (734) 975-7799; Web site: www.gsilumonics.com.

Figure 2. This chart allows an operator to click on an individual Pareto bar to reveal details on when the counted items occurred to help correlation with events such as shift changes.

Figure 3. An implementation of a Pareto chart that is particularly suited to component placement inspection before reflow is displayed. This chart uses cross-reference information to relate measurement data directly back to the placement equipment performance.

SPC Tools for Improving SMT Process Quality

The two SPC tools presented here are the Pareto chart and the X-Bar and R chart. When using control charts to monitor the process it is important to understand that they do not determine whether the product is good or bad, but rather whether the process is performing consistently. Inspection equipment should be able to make pass/fail judgements on each board using specification limits, as well as provide SPC tools to help prevent defects because of process variation. SPC charts (and in-line inspection equipment for that matter) should do more than simply indicate that the process is changing - they should also help point to the source of the change. Real-time SPC charts should be flexible enough to track individual devices, package types, or package classes such as ball grid arrays (BGA) or quad flat packs (QFP). X-Bar, R and Pareto charts can be used together to know when the process is changing. More importantly, they can also help explain why the process is changing.

X-Bar and R charts are a means to graphically monitor the process and determine if it is in control (experiencing only normal random variation within an acceptable range) or out of control (changing because of some external influence). The X-Bar and R chart pictured in Figure 1 automatically tests several control rules that are traditionally used to determine when a process is out of control. For example, one of the rules tests whether nine or more consecutive points have occurred on one side of the process mean, which can indicate a shift in the process. The chart pictured also tests an important statistic called process capability (Cpk). Cpk indicates whether the process, given its natural variation, is capable of producing product within the desired specification limits.

When an X-Bar and R chart fails, it may be difficult to understand why the variation in the process is occurring. The inspection equipment should be supplying measurement data that can help. Pareto charts such as the ones pictured in Figures 2 and 3 provide more concrete analysis of inspection data. These charts can provide the information needed in a nicely distilled format. A Pareto chart sorts and counts problem occurrences and presents them graphically to help isolate the most frequent problems. To be most useful for SPC analysis, a Pareto chart should allow counting of measurements that fall outside SPC control limits - not just measurements that fall outside specification limits.