How We Won the Regional Competition at Calsonic for Reflow Quality and Cost Improvements

By Rod Sampson, Process Engineer

Calsonic Kansei is a multinational automotive parts manufacturer with its own branded OEM products. Recently, the company was acquired by the private equity firm KKR, and it is poised for dynamic global outreach. The company’s main production focus is on quality, closely followed by cost effectiveness. This is no surprise for any company operating in the automotive industry and the highly competitive EMS business. At Calsonic, however, we view quality and cost effectiveness as mutually reinforcing in that poor quality is expensive, especially in the automotive segment where rework is quite undesirable for most, and altogether disallowed for certain processes. Implementing a solution for controlling and tracking our reflow process while optimizing productivity through maximum equipment utilization was the reason we won Calsonic’s regional Process/Quality improvement competition. This is our story.

I am responsible for the thermal processes that include reflow, wave, and selective wave soldering. For each new product introduction (NPI), we run Design of Experiment (DOE) to identify the “sweet spot” in the thermal process. This typically entails experimenting with machine settings that result in a process (profile) in the high, low, and center points of the process window. (Calsonic’s HQ provides us with the appropriate process window.) When the best process is identified, we lock it in place, then track and verify the process daily. While defects are rarely attributed to the reflow process, the challenges that are encountered when defects actually do occur are often more difficult to resolve. Quality is not something that should be weighted between the different processes, as weights are of little importance once an assembly has been rendered defective. Everyone should go about their business as if their respective processes are the most crucial/vital from a quality standpoint. If we all do that, we can minimize the impact of internal and external defects on our businesses, positively driving overall factory productivity.

The recent trend to aggressively reduce solder void has made process optimization and consistent soldering critical for our success. Previously this work was time consuming, and although the quality was acceptable, productivity and cost effectiveness suffered due to the necessary line stoppage to perform the work. It was challenging for us because we had numerous generations of reflow ovens and wave solder machines. Not only do their thermal properties differ, but the older machines made performance tracking difficult. In today’s age with virtually unlimited information available via a click, our method was outdated and inadequate. There had to be a better way.

We researched the market and found new technologies that seemed promising. One was a smart thermal profiler with predictive process optimization software that allowed us to input the process window target, and immediately identify the corresponding machine settings. This cut our process optimization time dramatically, and it allowed us to find more optimum process settings than in the past.

Previously, we had relied on technicians performing manual process verifications on each line multiple times per day. This was less than optimum because such manual verification is not consistent, and because it reduced our productivity by stopping the entire line for 10+ minutes every time a verification profile was run on a reflow oven. Additional verification events were prompted by yield issues, preventive maintenance work, and unscheduled line stoppage. (Our procedures require running a profile if the line stoppage exceeds four hours.) A secondary concern was that we had no thermal process information during the production between the twice daily manual profiles.

We decided to invest in KIC’s automatic profiling systems for reflow that use permanently embedded sensors in the oven to calculate each PCB’s profile on the fly during production. The automatic profiling systems were retrofittable on all our reflow ovens regardless of age, giving us consistent process tracking and traceability across all lines. To test the non-contact profile accuracy and repeatability, we forced changes in the oven while verifying that the system did indeed correctly identify the resulting out-of-spec event. I also found it interesting that we could stop a blower motor in the reflow oven without triggering an oven alarm or notification. This probably is because the temperature did not change. But the convection rate had changed, causing the profile on the PCBs to change, which is what determines our in-spec production.

The automatic system allowed us to control and document the reflow process. In addition to full thermal process traceability, the system will instantly alarm on an out-of-spec event. It also will provide SPC charting with Cpk limits that provide early warning on an out-of-control reflow process. Hence, we replaced manual verification profile spot checks with on-the-fly profiling of every PCB.

However, it was challenging to make use of the of thermal process data for trend analysis, yield troubleshooting, and other issues due to the sheer volume of data located on each reflow oven PC. One of our use cases would be triggered by a negative trend in our production line yield. In the past, it could take me more than a day (accounting for the normal interruptions in a workplace) to put together a trend analysis report to gain the appropriate insight to take corrective action on yield issues. When we stop the production line, we are keenly aware that every second of downtime is expensive. SPI and AOI inspection are very helpful. Sometimes the root cause of the problem is obvious, other times not. Fortunately, Vantage, a new factory-level database software, became available that made accessing the relevant data quick and convenient. Now I can identify the thermal process for a specific PCB or time period with a deteriorating production line yield. We will quickly rule the thermal process in or out, reducing troubleshooting time. And should reflow be the problem, which happens quite infrequently, then the automatic profiling system points us to where and what in the oven needs to be adjusted.

This progression made me realize that I had previously lacked timely reflow process data to make good decisions, and suddenly I had so much information that I was still not able to take corrective action quickly enough. The introduction of the Vantage factory-level database software changed all that by moving from data to insight. It is like looking for information online. You easily can capture massive volumes of data, but you need a strong browser or analytical tool to identify the relevant information or answer you need. You turn data into insight, and you want insight quickly.

A final dramatic use case, which improved our production line uptime by several hours for each event, was our occasional inability to run the profile verification due to broken thermocouples (TCs), damaged profile board or other. If such occurrences took place when the responsible process engineer was unavailable, for example on a night shift, then we shut the line down until a profile verification could be performed. Shutting the line down for several hours because a $6,000 profiler or a $30 TC broke may sound insane but, as stated above, Calsonic’s top priority is quality. We are, however, no longer held hostage by such events.

Every year, Calsonic conducts an internal competition on improvements in two main categories:
A. Quality
a. Improvements
b. Assurance
B. Cost reduction
a. Cost savings
b. Time savings

With the new smart profiler, automatic reflow process monitor, and factory-level database software, we automated quality assurance and we saved engineering and production time.

We won the regional Calsonic contest with these results:
1. Defects because of out-of-spec profiles all but disappeared
2. Automatic verification of process adherence
3. Effective void reduction
4. Near elimination of periodic profiling downtime led to an increase of production time
5. Set-up time for new products improved (NPI)
6. Line utilization gains of multiple minutes for each of the following events:
a. Preventive maintenance
b. Troubleshooting yield issues
c. Unscheduled stoppage
7. Full thermal process traceability for every produced PCB – customer satisfaction
8. Dramatic decrease of production stoppage due to inability to run manual profiles

The Future

As a company Calsonic has high aspirations of using technology to drive our core initiatives of quality and cost effectiveness. Personally, I envision a factory with smart machines and processes, with real-time dashboards connected to an MES system. Where manual and inconsistent tasks can be replaced with automation for better process control as well as improved quality and production. Eventually, we will be able to use data, insight and automation to run our factory more effectively. Perhaps there is another competition we can win soon…

For more information about Calsonic, contact the company at https://www.calsonickansei.co.jp/en/.

Reprinted with permission US Tech January 2019

https://kicthermal.com/wp-content/uploads/2018/12/Calsonic-How-we-won-the-competition.pdf