Thorough reliability analysis helps minimize process variability.

Capitalizing on the benefits of Lean manufacturing requires a companywide focus on elimination of non-value added activity and enhanced throughput. Reliability analysis resources can be great tools in supporting this effort. At EPIC Technologies, an in-house reliability laboratory supports new process definition and validation; new product process validation; and resolves internal, supplier and customer quality issues.

What are ways this activity supports process improvement, speeds new product introduction and resolves hard-to-identify quality issues in the production process?

The lab is co-managed by quality and process engineering to ensure eye-to-eye vision on support. A process engineer trained in reliability lab processes runs the lab. Data are collected in production per lab request; appropriate analysis is performed in the lab; reports are generated; appropriate stakeholder reviews are performed, and corrective actions are implemented.

New process definition and validation. EPIC is a long-term user of vapor phase technology. Detailed reliability analysis was used in both initial validation and subsequent fine-tuning of processes to better support the requirements of Pb-free processing and BGAs. Cross-section analysis using a scanning electron microscope is part of extensive failure analysis. Other key tools include shear analysis to analyze joint strength, 2-D and 3-D x-ray inspection, and strain gauge analysis to measure the strain in each process.

Study of the vapor phase process determined two preferred PCB finishes: immersion silver and electroless nickel immersion gold. Immersion silver has better solderability, but it is not as flat and tarnishes. ENIG is flatter and doesn’t tarnish. Oxidation is not a problem in the current process because suppliers are required to package PCBs in silver saver paper, and the Lean Kanban process used with the supply base ensures PCBs are consumed before adverse oxidation levels form.

Tests were conducted using VP reflow. Test vehicles were used with SnPb HASL or Pb-free immersion silver, immersion tin or ENIG surface finishes, as appropriate. Boards were populated with SnPb or Pb-free components, printed, assembled and soldered using standard reflow or VP production equipment. The solder pastes selected for testing included SnPb and Pb-free no-clean and water-soluble formulations. Assembled test boards were thermal-shocked between -45o and +125oC for 20 min. duration at each limit for 500, 1000 and 2000 cycles in EPIC’s FA lab. Other test boards were subjected to accelerated aging at 85oC and 85% relative humidity for 1000 hrs.

Vapor phase soldered boards were soldered in an EPM-IBL SLC500 vapor phase soldering chamber using Galden LS/230 perfluorinated heat transfer fluid. The vapor phase profiles developed provided a time above liquidus (TAL) of about 90 sec. and a maximum temperature of 230oC, a temperature governed by the vapor temperature. After a vapor phase profile is established, TAL can be modified to achieve any time required without exceeding the 230oC maximum temperature.

Visual inspection for solder balls, tombstones, bridging, voids and dewetting indicated no apparent difference between the two methods of solder joint creation. No tombstones were experienced on the test vehicles in either case – a positive result of proper pad, aperture and reflow profile design.

Visual inspection also indicated that while vapor phase solder joints performed well and microsection appearance was good, it might be a good idea to explore increasing the Pb-free TAL above the 60 to 90 sec. recommended by solder paste manufacturers to accommodate thorough heat transfer to larger components or clusters of large components. Another option is to increase the peak temperature of the vapor phase fluid. Larger thermal load components, especially in clusters, tended to retard the complete melting of Pb-free paste. It is more difficult to ensure good joints on components with high thermal mass in convection processing because while trying to achieve a sufficient TAL on larger components, smaller components in less populous areas may tend to overheat.

Another area of focus has been to match the size of the BGA to the BGA pad. This facilitates a better solder ball joint because variance in pad size and ball increases stress. Additionally, internal design guidelines oversize corner balls and pads on BGAs to provide more strength in the place likely to crack first.

Cracks, head-on-pillow (HOP) and knuckling are gross yet common defects that cause intermittent failures and are impossible to catch with AOI and difficult to detect with 3-D x-ray. Failure analysis can support process modification to prevent this.

NPI. The reliability analysis lab also supports NPI by providing data that can be used in product qualification. Typically customers provide two assemblies for this part of the qualification process.

One area of focus is BGAs, as solder joints aren’t visible without destructive testing. The lab cross-sections and verifies the BGA center and outside perimeters reflowed adequately.

Failure analysis also helps processes evolve to support emerging customer requirements. For instance, continued miniaturization is driving highly complex assembly such as package on package. In this situation, a BGA is placed and a second BGA is flux-dipped and placed on top of the first BGA. There is now a triple stack BGA, adding to the already challenging component ΔT. Vapor phase works well with those technologies because it reflows evenly at 230oC using SAC 305; however, other SAC formulations require higher temperatures. The lab is currently analyzing a specially formulated 235°C boiling point fluid in order to provide the best options to support these future customer needs.

Quality issues. Perhaps one of the areas in which laboratory service is most effective in contributing to enhanced throughput is its ability to help engineering determine the root cause of quality issues through failure analysis. The product and process validation techniques discussed earlier help optimize processes from the start, but even with a robust process, additional defects may occur. Often these defects arise not from the EMS process, but from issues in a supplier’s process or in handling at the customer. Failure analysis resources can help quickly identify the issue. The following examples illustrate how these defects can arise:

In one case, analysis indicated field failures were the result of the customer’s final assembly process (Figure 1). The analysis monitored the strain in each of the processes through ICT and functional test at the contractor, and strain in each of the customer’s processes. The result showed the root cause of the failures was a screw mounted next to the BGA during the customer’s final assembly process. Strain gauge analysis showed excessive and constant strain was applied to the BGA in question once the final assembly housing was fully installed, resulting in fractured solder joints and open circuit connections. The problem was quickly resolved, and it helped solidify the customer relationship because the recommended process modifications were based on quantitative analysis.


Supplier-related issues also can cause hard-to-identify defects. Some issues occur when suppliers are out of spec. For instance, in one case a customer’s product was leaking lithium fluid in the field. The problem was an issue in the battery supplier’s assembly process. Production operators were actually crimping and pinching the battery seal in final assembly (Figure 2). Once the failure analysis data were presented, the supplier took responsibility, and modified the batter assembly process, improving yield and reliability.


What started as an assembly workmanship concern in the eyes of the customer resulted in a supply base issue; an issue where root cause was identified and permanent corrective action was taken. This type of customer/supplier/manufacturer problem-solving partnership would never have been possible without the use of undisputable quantitative analysis provided through reliability laboratory services.

While reliability analysis isn’t traditionally thought of in terms of Lean manufacturing, these examples demonstrate how its use helps minimize process variability, development and speed manufacturing and quality issue resolution. The net effect is improved customer quality, enhanced manufacturing and resource efficiency, leading to improved throughput.

References

1. Chris Munroe, “The Advantages of Vapor Phase Processing in RoHS-Compliant Assembly,” SMTA International Proceedings, August 2008.

Ryan Wooten is process engineering manager at EPIC Technologies (epictech.com); ryan.wooten@epictech.com.