-
- News
- Books
Featured Books
- smt007 Magazine
Latest Issues
Current IssueBox Build
One trend is to add box build and final assembly to your product offering. In this issue, we explore the opportunities and risks of adding system assembly to your service portfolio.
IPC APEX EXPO 2024 Pre-show
This month’s issue devotes its pages to a comprehensive preview of the IPC APEX EXPO 2024 event. Whether your role is technical or business, if you're new-to-the-industry or seasoned veteran, you'll find value throughout this program.
Boost Your Sales
Every part of your business can be evaluated as a process, including your sales funnel. Optimizing your selling process requires a coordinated effort between marketing and sales. In this issue, industry experts in marketing and sales offer their best advice on how to boost your sales efforts.
- Articles
- Columns
Search Console
- Links
- Events
||| MENU - smt007 Magazine
Implementing AOI in a CM Environment
December 31, 1969 |Estimated reading time: 6 minutes
Adding AOI to a typical manufacturing line can lead to improved first-pass yield at ICT.
By Shane Downing and Mark Owen
The SMT contract manufacturing (CM) world is undergoing dramatic changes. Many companies are growing more than 40 percent per year, as OEM companies in the telecommunication, computer and other industries divest themselves of assembly operations. These changes have impacts on the SMT line:
- New lines added, moved, modified and upgraded at a faster rate
- Integration of expectations, standards and quality procedures
- A flood of high-technology products and new component types entering the factory.
One CM's strategy* to implement automated optical spection (AOI) was aimed at improving SMT line process control, ensuring high in-circuit test (ICT) yields and establishing measurable benchmarks across all SMT lines. At this facility, the product mix is approximately 50 percent workstation boards, 30 percent server boards and 20 percent telecommunications.
Process Control StrategyThe process control strategy for this facility was developed by evaluating the following metrics:
- Component types presently used and expected in the future
- Defect types and quantity
- Line cycle times
- Maximum board dimensions
- Value analysis to determine where improved control would maximize benefits.
Component mix ranges from 0402s to 0.020" quad flat packs (QFP), and array-area devices like ball grid arrays (BGA) and chip scale packages (CSP) are growing in volume. At the time of purchase, it was decided that 0201s would not be in volume use for several years, and less than 0.020"-pitch devices would be supplied in area-array formats. A very low percentage of leaded through-hole components are used in the product mix.
Previously, a 3-D paste inspection system was added in one line. While this had some positive effects, first-pass yield at ICT still was only 75 percent. An analysis of these defects showed that pre-reflow component placement and hand-insertion processes caused most failures.
The placement rate of the fastest line is 60,000 components per hour (cph) and 84 percent are passive devices. Each line has a changeover every one to three days and sees about five new designs per year.
Based on this information, it was determined that component placement inspection would be implemented, and the machine must satisfy the following requirements:
- 100 percent inspection speed faster than 90 seconds per assembly
- Must handle 0402s, 0.020"-pitch QFPs and BGAs
- Must provide information sufficient for process troubleshooting
- Minimum setup time for a new design on changeover.
A system** was chosen based on these criteria, and although there were no solid plans to use the system for a measurement, the potential was understood that it would eventually be an integral part of process quality.
Choosing where to install AOI is company and site specific, depending on existing process controls, component types in use, line throughput and the production line problems. At this CM, the choice was to add in-line component measurement pre-reflow because:
- It allows defect detection at an easy repair point
- Component placement had been determined to be the highest cause of random failures
- Pre-reflow systems are easier to program and have lower false-fail rates
- The assembly mix has a high number of BGAs.
Implementation IssuesThe inspection system was implemented in several lines. Installing a rework station after placement allows the insertion of manual components and component verification.
It took seven days to reconfigure the line to add the inspection and rework stations, train the engineers and operators, and get the first boards operational. Additional operator training (providing the information necessary to load programs and inspect failures) took one to two hours. In this instance, the general place operator runs the inspection process.
Typical production changeover time between products is three minutes. Typical setup for a new board is four hours when a new component database needs to be defined. Less than 30 minutes is required for a new board with an established component database where most part numbers and shape codes are already in place.
Figure 1. First-pass yield at ICT improved from 75 to 93 percent and has remained there since adding in-line component placement measurement.
The first problem encountered after the installation of the original system was its inability to translate component placement data from one pick-and-place machine type. This was resolved by developing a temporary converter until one was available from a third-party software supplier. The second and final implementation difficulty occurred after multiple systems were installed, and new products and components were added. False error detection became unacceptably high (greater than 1 percent) because of dull leaded components and bright pads with cutback paste. The following software changes were made to reduce the false-fail rate:
1. Separate databases per product allowed each board to have unique component definitions.
2. Introduction of a multi-variable classifier to better define component images. This tool better-defined component images and allowed false fails to be trained out over time.
After the implementation of these above changes, the false-failure rate decreased to less than 0.5 percent. Currently, the system detects errors at the rate of 375 ppm real vs. 230 ppm false failures.
ResultsImplementation of AOI tools for component placement inspection was used for several specific processes, including:Identifying Process Problems:
- Missing and misplaced components: By detecting this pre-reflow and manually correcting them, a significant increase in first-pass yield at ICT was achieved (Figure 1). Passive components (0805s, 0603s) appear to have the highest defect rate. The inspection system was credited for driving first-pass yields at ICT from less than 75 to greater than 93 percent.
- Board supports and bent nozzles: The inspection system was used to link the root cause for missing parts to items such as misplaced board supports and bent nozzles on pick-and-place machines.
Identifying Process Improvements:
- Pick-and-place machine X/Y offsets: Based on measurement information obtained from the inspection system, pick-and-place machines were recalibrated to bring placements closer to computer-aided design (CAD) data (Figure 2).
- Moving hand-load operations pre-AOI: It was determined that pre-reflowed components were bumped during hand-load operations on certain assemblies. At the time of this discovery, hand-load was located post-component inspection. Based on this discovery, hand-load moved to pre-component inspection so any components bumped during hand-load were flagged as errors prior to being reflowed.
Figure 2. Example of measurement information available from inspection system.
Establishing Process Metrics:
- Two lines have an inspection/repair combination just prior to the reflow process. Using component failure information obtained from the repair station, operators track missing, offset and polarity defects. If quality levels drop below 90 percent or if three errors of the same type are found on consecutive boards, the process is halted and corrective action is taken.
- Prior to moving each of the eight SMT lines to a new facility, a capability baseline for each was established. After the equipment was moved, the same process was followed to ensure each line met this baseline. The process to determine pre- and post-move pick-and-place capability was as follows: 0805 passives, 0.050"-pitch 20 pin plastic leaded chip carriers (PLCC), and 0.020"- pitch 240-pin QFPs were placed on 10 qualification boards; each board was inspected by the measurement system to obtain X/Y and Theta placement data; Cp and CpK numbers were then generated for each machine using software provided by the manufacturer. This process is also used when new lines of placement equipment are added.
ConclusionSince 1997, three lines have been fitted with component placement inspection at this CM. These tools found placement defects before reflow, helping to improve first-pass yield at ICT from approximately 75 to greater than 93 percent. Process control measurements and machine characterizations have become essential tools for determining placement machine calibration requirements, comparing multiple machines from the same supplier or the performance of different lines, and quantifying improvement efforts. Using intelligent vision algorithms has significantly reduced false calls because of dull leads, gold plated traces, cut-back paste and variations between components from different suppliers.
*Celestica's Ft. Collins, Colo., facility.**GS-1
SHANE DOWNING and MARK OWEN may be contacted at MV Technology Ltd., Unit 3, Enterprise Centre, Pearse Street, Dublin 2, Ireland; +353 1 671 81 77; Fax: +353 1 671 84 70.