Estimated reading time: 11 minutes

SMT Solver: How to Audit OEM-EMS Assembly Capability, Part 2

As I mentioned in Part 1 of this column series, during the past two decades, there has been a tremendous increase in outsourcing by OEMs to EMS companies, which also results in a decrease in yield. I further discussed the overview of an audit process to assess the manufacturing capability of OEMs or EMS companies. In this column, I will focus on the technology and manufacturing capabilities of the supplier.

Even though I am dividing the questions into various areas—namely technology, manufacturing, quality, and RoHS compliance—there are going to be overlaps. However, it is not important which category these questions fall into, but they should be asked. They are intended to help generate questions that are relevant to your product and the manufacturing site you plan to audit.

Manufacturing Questions
There are some business-related questions aimed at determining the supplier's financial stability, long-term viability as a business, pricing policy, and quality standard to be used. They should be asked. You should also ask about the production capacity of the line and the percentage of that capacity available for your products to see if there is a match. I am not covering these business areas in this series of columns; instead, I’ll focus only on the technical areas.

You can begin by asking about the types of components used in various products being built by the company today and the level of defects (PPMO and first-pass % yield) in those products. The answer to this one question will give a good general overview of the assembler’s capability.

You can ask the supplier to check the types of packages and their pitch being used in the products they are currently building. Here are some examples, but the basic idea is to know the capability of the supplier to handle the largest and smallest I/O counts and lowest pitch to make sure types of components you have in your products are being built by the supplier:

• Types of through-hole components being used
• Smallest components (0402, 0201, and 1005)
• Resistor networks with fine-pitch pads (0.4-mm pitch)
• BTCs, such as QFN, DFN, LGA
• QFP with 0.4-mm and 0.3-mm pitches
• CSP/BGA with 0.5-mm pitch
• CSP/BGA/fine-pitch with lower than 0.5-mm pitch
• Maximum BGA I/O count
• Package-on-package (PoP)
• Others (please list)

The types of packages and their pitches play a key role in the level of defects. For example, as I have discussed in my previous columns, here is a brief summary of defects you can expect for different types of packages:

• Plated through-hole (PTH): 4,000 PPM
• Gull-wing: 1,400 PPM
• Chips, BGA, J-lead: Around 600 PPM
• Average of all types: 1,079 PPM

What is noteworthy in this study is that through-hole components cause the highest level of defects. It should not be a surprise since the most common process for soldering through-hole components is wave soldering, which has too many variables to control. Since through-hole is not going away for some time to come, even though their number has dwindled drastically, you or your supplier may want to consider automated selective soldering for PTH to reduce PTH defects.

In the same study, the author found that the type of lead and its pitch plays an important role in defects:

• 16-mil pitch (0.4 mm): 13,088 PPM
• 20-mil pitch (0.5 mm): 1,878 PPM
• 25-mil pitch: 950 PPM
• 50-mil pitch: 650 PPM

When the pitches get below 0.5 mm, the defect level really skyrockets. If you cannot avoid 0.4-mm pitch, you really need to focus on the manufacturing capability of your supplier to successfully deal with fragile leads of ultrafine-pitch packages. Put another way, if your product contains pitches below 0.5 mm, there are very few companies that can build them with good quality on a consistent basis. Thus, your audit process must be a lot more rigorous if you need to assemble ultrafine-pitch components.

There is no such thing as a perfect PCB surface finish. All surface finishes—such as HASL, OSP, ENIG, immersion silver, and immersion tin—have their pros and cons. In most cases, the assembler does not make their own PCBs; they order them from a PCB supplier. An assembler who is a turnkey supplier is responsible for selecting their own PCB supplier. On the other hand, if you specify the surface finish, you should ask the same questions from the PCB supplier.

Here are four examples of questions that should be asked about different surface finishes being used on your product:

1. ENIG surface finish: Did you ever have black pad incidence? How was it resolved? Very few suppliers are willing to admit the problem, but black pad is a potential problem with ENIG, and you should ask about it.
2. Immersion silver: Have you ever had champagne micro-void incidence? If so, how was that resolved?
3. OSP: Have you ever had BGA ball drop incidence or any issues with via fill during wave soldering of mixed assembly? Did you have to use a more active flux or nitrogen to achieve 100% via fill? Even though only 75% via fill is required, if you always get that 75% minimum, it is not acceptable quality.
4. HASL: Since the inherent problem with this finish is an uneven surface finish, it is really worth asking whether the supplier has successfully used HASL for BTC, BGA, or fine-pitch packages. Very few people can make that claim, and that was the reason for moving away from HASL in the first place. It was not too long ago that the predominant surface finish was HASL. In addition to getting uniform solder coating with HASL, board warpage may also be an issue with HASL. For most companies using finer pitches, BGAs, and BTCs, HASL is really not an option.

Because of availability constraints, many companies end up using both tin-lead and lead-free components on the same board. Such assemblies fall into what is commonly known as forward or backward compatibility scenarios. The key concern is the selection of the right peak reflow temperature since the tin-lead and lead-free components require very different peak reflow temperatures for proper reflow.

Key questions to ask are as follows:

• How does the assembler deal with tin-lead and lead-free BGAs on the same board?
• What peak temperature is used to prevent overheating of tin-lead BGAs/QFPs without compromising the proper reflow of lead-free BGAs?
• What reflow peak and TAL do you use when most of your components are tin-lead, but the BGAs are lead-free?
• Does the supplier know the difference between TAL and true TAL? (TAL is time above liquids even if only one component sees that temperature. True TAL, on the other hand, is time that all components being monitored see that temperature. It makes a big difference when you measure the temperature of BGAs—especially the temperatures of inner and outer BGA balls.)

Handling and control of moisture-sensitive components and tracking of their exposure times is another issue that is worth digging into since very few companies do it correctly:

• What is the handling procedure to prevent baking (when exposure time has expired) of moisture-sensitive BGAs?
• How many times do you bake moisture-sensitive components (only once is allowed for some MSL levels)?

It is not easy to detect the seriousness of the defect in a company during the course of a short audit. But finding the ratio of bridging to opens over the last six months can give you a good idea about the extent of field failures you may see. About six times more shorts than opens are a good sign since there is practically zero chance that shorts would escape any inspection or test. Opens, on the other hand, can easily escape inspection and test and will come back to haunt you after a relatively short time in the field. Don’t be surprised at all if you have a very difficult time finding any supplier that can meet these criteria, but be surprised (and happy) if you find one with more shorts than opens in their products.

It is also important to ask about the test strategy (ICT, flying probe, functional, etc.) used on the product—especially when you see a large bone pile on the shop floor. A large bone pile is common in companies that do not use ICT and have high defect rates.

Technology Questions
The types of machines used by a company are important. However, the existence of detailed and formal design and process documents and extensive training programs for operators, technicians, and engineers are the key elements of a strong SMT infrastructure of a company. You should ask about the number of engineers working on manufacturing process development and on the production floor, and their qualifications. And the questions should be intended to assess their understanding of the technology.

For example, you should ask about the properties of solder paste (compositions, metal content, particle size range, etc.) to see if they have been selected for a reason:

• What is the solder application method (stencil or screen)?
• Why was this method chosen?
• What is the paste deposit thickness?
• Does this thickness requirement change if fine-pitch is used on the same board?
• What is the approach (differential stencil thickness versus micro-modification of stencil aperture) for applying the paste on a board with standard surface mount and fine-pitch?
• Have they looked into the implications of each approach?
• Do they get heel fillet in fine-pitch without getting insufficient fillets in standard components?
• What is the reflow soldering method used (vapor phase, infrared, convection, or combinations thereof)?
• What are typical thermal profiles on the board surface and at a solder joint?
• Do they develop a unique profile for each board?
• How and where are the thermocouples attached?
• Are they attached on the surface or on the solder joints?
• Is the board drilled to attach thermocouples to the BGA balls on the outer and inner layers?
• If not, how do they know if the BGA balls really reflow?
• Have the reflow processes been compared for manufacturing yields?

In developing a profile, there is an inherent conflict between the type of profile you need for a BGA (shorter soak time) and what you need for BTC (longer soak time). What does the supplier do when both are on the same board, which is a very likely scenario? Developing a reflow profile for electronic assemblies is like trying to figure out the time and temperature (bake profile) to bake turkey, chicken, and shrimp for the same length of time at the same temperature in the same oven without undercooking the turkey or overcooking the shrimp.

Your questions should focus on assessing the inherent technical capability of personnel to resolve complicated technical issues that they need to deal with all the time in a complicated assembly:

• What cleaning method and solvents are used?
• How is the cleanliness of product boards monitored?
• What sort of repair/rework equipment is used?
• What are the thermal profiles for repairing each surface mount device type?

To determine whether the company has any experience or plans to develop capabilities for newer technologies such as BTC, BGA, and ultrafine-pitch packages (if your product requires them now or in the future), you should ask detailed questions related to that technology. For example:

• Are you planning to achieve toe/side fillets in BTCs, and if so, how?
• How do you ensure you get sufficient paste thickness to prevent potential opens without getting too many voids?
• How do you prevent paste dripping into via holes in thermal pads?
• What do you see as the major concerns in PBGA, and how do they compare to PQFP?
• What is their experience with head-in-pillow (HIP)?
• If you are doing all you can about paste height, reflow profile, do you think the component supplier is responsible for HIP?
• Do you prefer copper-defined pad or solder mask-defined pad, and if so, why?
• What do you think of the approach used by some companies where solder mask is sprayed up to the edge of the pad with zero clearance?
• Do you prefer via in pad or tear-drop design, and if so, why?
• Did you conduct any evaluation on PCB routing for different pitch, pad sizes, line width/space, etc., and their impact on layer count?
• Did you experiment with different pad sizes and their impact on process yield?

Conclusion
The idea behind the questions in this column is to establish whether the company has done extensive process evaluation and whether they understand the importance of critical materials and process variables on product quality and reliability. It will be obvious within a very short time during the visit if there is someone in the company who understands these issues or if they are just following the recommendations of component and material suppliers. How many component or material suppliers do you know who build PCBAs? How useful would their recommendations be?

I should also note that the intent of these questions is not to dictate the process to the company but to assess their understanding and capability. You should focus mainly on the end requirements. Let the assembler worry about how best to meet those requirements. You are simply trying to establish whether they can meet your requirements.

In my next column, I will conclude this series with questions on quality and RoHS compliance.

Ray Prasad is the president of Ray Prasad Consultancy Group and author of the textbook Surface Mount Technology: Principles and Practice. Prasad is also an inductee to the IPC Hall of Fame—the highest honor in the electronics industry—and has decades of experience in all areas of SMT, including his leadership roles implementing SMT at Boeing and Intel; helping OEM and EMS clients across the globe set up strong, internal, self-sustaining SMT infrastructure; and teaching on-site, in-depth SMT classes. He can be reached at smtsolver@rayprasasd.com and regularly offers in-depth SMT classes. Details about classes can be found at rayprasad.com.

This column originally appeared in the November 2020 issue of SMT007 Magazine.