-
- News
- Books
Featured Books
- smt007 Magazine
Latest Issues
Current IssueIPC APEX EXPO 2025: A Preview
It’s that time again. If you’re going to Anaheim for IPC APEX EXPO 2025, we’ll see you there. In the meantime, consider this issue of SMT007 Magazine to be your golden ticket to planning the show.
Technical Resources
Key industry organizations–all with knowledge sharing as a part of their mission–share their technical repositories in this issue of SMT007 Magazine. Where can you find information critical to your work? Odds are, right here.
The Path Ahead
What are you paying the most attention to as we enter 2025? Find out what we learned when we asked that question. Join us as we explore five main themes in the new year.
- Articles
- Columns
Search Console
- Links
- Media kit
||| MENU - smt007 Magazine
Estimated reading time: 8 minutes
Contact Columnist Form
SMT Solver: Major Drivers for Improving Yield and Reducing Cost
No matter how long you have been in the SMT industry, there is always something new to learn because SMT is constantly evolving even though it is a mature technology. For example, we have been manufacturing SMT products in high volume for almost four decades, but less than 10% of companies have a first pass yield of more than 90%. In other words, 90% of companies are conducting too much rework.
In some of my previous columns, I have gone into detail on the types of defects that are predominant. There are many reasons we worry about defects, but the key reasons are increased cost and reduced reliability. These defects must be reworked so that you can meet the delivery requirement without too much delay. Even if you can minimize the delay in shipping the final product, if you must do rework, there is no way to prevent reduction in reliability. Why? Each time you heat the solder joint, you increase the thickness of the intermetallic layer. The IMC is necessary to achieve a reliable bond between the component and board, but too much IMC is not good. It is brittle and if it is too thick you have potentially early failures in the field. There is nothing worse than field returns for any company since they are not only expensive to repair, but they can also adversely impact a company’s reputation.
With the advent of fine- and ultra-fine-pitch, high-pin-count BGAs, 0402, 0201 and 01005 resistors and capacitors, as well as the widespread use of no-clean flux, yield problems are getting worse. With widespread use of lead-free, the yield problems are compounded. When yield problems persist, most people blame manufacturing. This is unfair and prevents companies from implementing the necessary corrective actions.
Who in the company is responsible for the defect? If you really think about it, no matter what your job title is or which department you work in—engineering, manufacturing, quality, purchasing, or management—what you have in common is the responsibility for quality and cost. You affect quality and reliability in different ways. No one group can ever solve this problem alone. If you take a 50,000-foot view of the problem, there are three major areas that affect quality: design for manufacturing, quality of incoming materials, and manufacturing.
How the board is designed, and what components are selected, impact cost and quality. This is mostly the responsibility of the designer. Just about everything you need for manufacturing is purchased from outside—components, boards, and materials such as solder paste and flux. The person responsible for vendor selection is generally the purchasing manager. Yes, defects are caused in manufacturing but there is nothing they can do about poor design or bad quality of incoming materials. So, let us look in more detail at these three areas: design for manufacturing, incoming material quality, and manufacturing.
Design for Manufacturability
Keep in mind that DFM is a key driver, if not the most important, of manufacturing yield. However, few circuit and board designers understand manufacturing processes. A DFM document must be company specific. Using an industry standard such as IPC-7531 (formerly known as IPC-SM-782 when I initially chaired this committee in the mid-‘80s), is a good place to start. Some major items that should be included in a DFM for SMT products are:
- Establish design rules and guidelines while emphasizing the importance of differences between them
- Component selection criteria, including consolidation of parts lists to reduce redundancy and eliminate obsolete parts
- Paneling considerations
- Fiducial requirements
- Land-pattern design
- Solder-mask consideration
- Via-hole location
- Design for test
- Anything unique to your design
With widespread use of high-pin-count BGAs that cannot be inspected visually, sufficient test coverage for in-circuit test (ICT) should be seriously considered. Keep in mind that no inspection method is perfect. The only way to prevent defects from escaping to the field is to rely on overlapping test and inspection methods. Once a DFM document developed by a well-trained team is finalized and released, the possibility of DFM violation generally does not arise.
Creating a DFM document is not easy; it will, however, correct problems at the source and prevent their recurrence. This is critical in an environment where essentially all manufacturing is being outsourced or sent offshore.
Incoming Materials Quality
No matter what components, boards, solder paste, flux, etc., the designer and manufacturing people selected, the quality of incoming material is controlled by the vendors who supply them. Who is responsible for vendor selection? The purchasing guy who used low price as the criteria for placing the order? When it comes to making the decision about vendor selection, how much focus is given on price vs. quality and the total cost?
“Garbage in, garbage out” could not be truer than in the assembly of SMT components where pitches are shrinking and process windows tightening. As a result, there is no way to improve manufacturing yield if the boards and components have poor solderability or unacceptable co-planarity. Referencing industry standards such as J-STD-002/003 is a good idea. Take another example. BTC is now a very widely used component. It is a very good component with many good features, such as excellent electrical and thermal performance, size, and weight, etc. But the one important feature of this package is that it is the cheapest package you can buy. As a rule of thumb, if you can buy a component package that costs one penny per lead (i.e., a 100-pin packaging costing a dollar), you have an inexpensive package. A BTC costs less than half a penny per lead–no wonder it is a very popular package. But when it comes to manufacturing, you need a perfectly flat package, and a totally perfect flat board. Since there is no lead or ball in this package, packages and boards must be flat to ensure a good connection. To compensate for flatness, you cannot print excess solder which will cause floating and excessive voids. And you cannot print very thin paste thickness since the potential for opens will go up if either the package or the board is warped. Plus, the ends of the BTC terminations are not solderable since they are exposed copper. So, if you consider these manufacturing challenges, the package is not as cheap as one might think. I suggest you refer to the just-released IPC 7093B (I chair the committee), for a deep understanding of the complexities involved in design and manufacturing with BTCs.
Manufacturing Processes
The key responsibility of manufacturing is to use process control for all the manufacturing processes. However, no matter how good the process controls are, or how well they program their machines to dispense or print paste or use the right profile, they cannot eliminate design- and material-related defects. Yes, they can compensate for some of the issues with some process changes but there is a limit.
How should one identify key manufacturing process issues? For manufacturing there are a lot of challenges. For example, the equipment must be characterized thoroughly. This can be defined as understanding all parameters that affect the equipment’s performance. A good understanding of all the key parameters that affect quality will take a lot of time and effort. Large companies can afford to assign many engineers to characterize the process. In small companies, “learn as you go” is the common motto since they don’t have the resources. Vendors may say it is easy; it is not. For example, if you look at the data sheets of various solder pastes from different suppliers, what you will see is that just about any profile will work with their paste. That is simply not so. While you do consider their recommendations, there is no substitute for developing a unique profile for each product.
How should you proceed? First, characterize the process, then document the details of equipment- and non-equipment-dependent variables that control yield. There are some misconceptions that if you use a certain paste all your defects will go away, or if you buy a particular convection oven, there’s no need to develop a unique profile for each product. This is not true, as each board has a different thermal mass.
With the widespread use of BTC, BGA, and fine pitch, the challenges for manufacturing are only increasing. We briefly mentioned BTC issues in the incoming material quality section earlier. Dealing with BGAs is no walk in the park. One of the biggest challenges in BGAs has been the head-on-pillow defect. If you scan the literature, you will find multiple causes of head-on-pillow such as design and processes, paste, profile, etc. Some of it is true. But the inherent cause of head-on-pillow is the warpage of the package that very few component suppliers will admit to. This is a very involved subject that will require multiple columns to address. Suffice to say that even when you use the best possible paste and best profile and everything else you do in manufacturing, if the package is warped, you will not get rid of head-on-pillow. You can minimize the problem, but to eliminate it, the real solution is to use a package that does not exhibit unacceptable warpage at soldering temperatures (not room temperature).
For an in-depth look at BGA design and assembly challenges, look at IPC-7095, which I also currently chair.
Conclusion
In addition to having the right design, quality incoming materials, and good manufacturing capabilities, you need in-house detailed DFM and manufacturing process recipes; well trained personnel at all levels, including operators and technicians on the manufacturing floor; and process, design and quality engineers. Very few companies have detailed in-house DFM and process documents, and training budget is the first thing that gets cut at a lot of companies. Just keep in mind that no one gets up in the morning and says, “I am going to screw up three things today at work.” They are all trying to do things to the best of their ability. Who is responsible for the in-house documentation and training? The top management. The buck for quality, reliability, and cost stops at the boss running the show.
This column originally appeared in the April 2021 issue of SMT007 Magazine.
More Columns from SMT Solver
SMT Solver: Flux and Cleaning—How Clean Is Clean? Part 2SMT Solver: Flux and Cleaning—How Clean Is Clean? Part 1
SMT Solver: Optimizing Your Manufacturing Operations
SMT Solver: How to Audit OEM-EMS Assembly Capability, Part 3
SMT Solver: How to Audit OEM-EMS Assembly Capability, Part 2
SMT Solver: How to Audit OEM-EMS Assembly Capability, Part 1
SMT Solver: Industrial Revolution 4.0—Hype, Hope, or Reality?
SMT Solver: Dealing With Package Parasitics