01 February 2005

Plant floor visibility

By Sean McCloskey

Turning information into knowledge improves manufacturing process.

Shaun McGinnis knows the value of being able to tell you where a specific resistor is located—even if that resistor is on a single printed circuit board (PCB), in a data storage system the size of a refrigerator that has more than 400 PCB's and drives. McGinnis implemented a system that can tell you what time an operator installed any assembly in the unit, by which operator, all test history of the assembly, and finally, automatically update a manufacturing requirements plan (MRP) system. The next challenge is to track all assembly components, like a resistor, from the PCB serial number to the lot code.

Hitachi working to increase level of shop-floor detail.

More importantly, McGinnis knows how to get to this level of shop-floor detail in a way that produces a bottom-line benefit for his organization—Hitachi Computer Products (America), Inc. in Norman, Okla.

As a project manager, McGinnis along with a team of coworkers were responsible for finding a way to achieve detailed visibility on what was happening on the plant floor in order to help his company—a producer of large-scale data storage systems—effectively expand its professional and engineering contract manufacturing services.

Searching for solutions that provide more effective tools for tracking products, including quality and work in process (WIP) data, from the beginning of the manufacturing process through shipping has been a constant challenge for most manufacturers. Conventional wisdom was to invest in a manufacturing execution system (MES); however, over the past few years, Hitachi officials wondered if traditional MES solutions were able to deliver.

MES confusion

The reason for their confusion was they tried to get their arms around what MES is and is not. Dominated by legacy technology, point solutions, and small, unstable vendors, the MES marketplace has been anything but clear about what the solution offers manufacturers, especially those engaged in Six Sigma or lean manufacturing projects.

"It was only after we took a serious look at what solutions were available to help us improve shop floor visibility that we realized how the MES marketplace was transforming itself," McGinnis said. "Our research kept pointing us back to a handful of MES providers which were leading the way in moving the industry into an enterprise-wide solution that could provide the level of information needed to reduce operational costs, improve production processes, and, for the first time, give a company manufacturing visibility across a global enterprise from start to finish."

Beyond getting to the shop floor information, they needed to provide the company with more detailed reporting on defects per million, first pass yields, and yields. Hitachi also had to ensure this extra layer of data collection and evaluation did not impact lead time.

Hitachi had just completed a multi-year effort of implementing a combination of lean manufacturing and Six Sigma principles. Joe Soliz, head of test engineering for Hitachi and a Six Sigma black belt, realized the need for more information but was not willing to compromise the successes of the lead-time initiative he had just completed.

"We worked very hard to squeeze a lot of inefficiencies out of our manufacturing process through a lean Six Sigma approach that helped us accurately capture downtime that was unnecessarily extending our lead time," Soliz said. "As a result, we were able to dramatically cut our lead time from 12 days to around 4.5 days. This produced a significant savings and was something that could not be sacrificed."

Understanding these constraints, McGinnis led a cross-functional team in the search for a solution that would help Hitachi collect more data from its shop floor in order to effectively meet the needs of its customers and the growing government and environmental regulations. However, all the while, the team knew in collecting this data that it could not add additional employees nor negatively impact the lead time of the computer systems being assembled.

Given the full support of the company resources to accomplish its mission, McGinnis' team carefully reviewed what they needed, and, receiving in-depth guidance and input from Soliz's quality assurance team, the project team decided to go a different route.

Built upon a series of flexible components, they looked for a MES provider that combined advanced industry methodologies and the latest in Java 2 Platform, Enterprise Edition (J2EE) technology to define, track, and report on real-time information from a user's shop floor. (J2EE defines the standard for developing component-based multi-tier enterprise applications.) The result enables manufacturers to build upon Six Sigma and lean manufacturing programs to manage and control production with real-time information from a shop floor.

With an underlying technology that permitted it to connect to Hitachi's ERP, the team needed their MES provider to deliver on the promise of providing technology that filled the gap between business and control systems for manufacturers.

Implementation

The Hitachi project team created five sub-projects in order to best execute on its mission of delivering shop floor details without having an impact on lead-time. The five sub-projects were: genealogy, configuration, manufacture scan, process control, and revision controls.

However, in order to initiate the first phase of implementation, the team had to establish informational foundation levels where they could create layered analyses and reports. This was a key objective because the tracking and collection of the wrong data could adversely have an impact on the lead-time.

"The learning curve on collecting very detailed data can be very costly if you don't carefully think through exactly what your organization needs," McGinnis said. "Information for the sake of information will kill you and your results. Knowing what information can be quickly turned into actionable knowledge that can help your manufacturing process is where the real ROI on a project like this is often made or lost."

Working with their provider to closely understand what data and information could be provided, the team defined a list of key decision points needed before they initiated the phased implementation. Those defined decision points included the following:

• Items—determine the items they would need to track (Quality & WIP).
• Operation Names—define the names of the operations.
• Production Routings—define the production routings.
• NC Routings—define the rework/repair routings for failed items.
• Next Numbers—define serial number scheme for items.
• NC Codes—determine NC Codes and their hierarchy.

Following that, the project team led the development of a model system for pre-implementation testing to ensure it supported a low-volume/high-mix operation with complicated processes and unique features. The base product was installed in August 2002.

The five-phase implementation schedule was:

Phase I

• The creation of Shop Orders and the generation and introduction of the provider into the production process
• The collection of visual inspection data in the PCB shops
• The collection of test (pass/fail) data in the PCB Test area

Phase II

• The collection of test data (pass/fail) data in the system test area consisting of 150 different tests in four functional areas

Phase III

• The implementation of the PCA cartridge dashboard in the PCB test troubleshoot and repair area

Phase IV

• Establishing the automatic data feed to the provider from the HP3070 ICT testers

Phase V

• Currently in progress: Genealogy project (traceability)

Completed: Configuration project (using the provider to determine the configuration of a unit in the unit assembly area—also used to configure a unit to the final shipping configuration)

McGinnis' group knew the successful transition to a more proactive shop management system was going to be demanding on all parties involved. For that reason, training was factored into the program. Both CRP and on-line training was performed before and during the implementation phases. Ensuring assemblers and managers were educated on how the systems worked and what could be accomplished with it was vital.

Many times this retraining involved teaching a highly skilled and experienced workforce new ways of doing jobs they have done differently for years—a task not usually met with ease.

"In one case, we had to re-engineer the way the WIP tracking process was performed so serial numbers of sub-assembly components could be tied to a specific shop order," McGinnis said. "This was a major change, but because our management team did such a terrific job of getting behind this project, everyone knew that being able to track to this level of detail was needed to ensure our competitiveness as a contract manufacturer. The company-wide understanding of our team's work meant we met with almost no resistance to change."

Plugging the holes

The implementation occurred in phases, beginning with replacing the existing data collection points and the collection of the most business-critical data. Hitachi's previous data collection system remained active until all training was complete and all the 'gaps' filled.

As Hitachi's data collection increases in granularity, it becomes more useful. The extracted data is already seeing use in several projects and activities.

Soliz's Quality Improvement Teams (QIT) have quickly integrated the system's capabilities into their lean and Six Sigma efforts, using MES data and reports regularly.

The group now analyzes printed circuit board visual inspection data to improve overall quality. As a result of real-time shop floor visibility, the implementation quality in those areas has improved drastically. The "no defect ratio" has met or beat the company's goal every month since bringing the new technology online in January 2002. This key performance indicator is a metric that corporate management uses to evaluate Hitachi.

"Being able to get to this level of data without extending our lead time gives our lean and Six Sigma efforts an even greater chance of improving our bottom line," Soliz said. "We are just starting to get at information that we were never able to track before, and, as a result, Hitachi Computer Products (America) has consistently improved its quality measures."

Soliz and McGinnis expect Hitachi to experience similar successes as the project continues to move forward. The Electronic Mechanical Assembly group just went live with the new technology. Three months ago, paper files would have had to be pulled and examined to know what specific hard drive went in what computer rack. It would have taken two to three hours of data entry per computer cabinet to collect and record the information needed for accurate tracking to the specific part level. Now, this process takes less than 30 minutes to perform.

Soon, McGinnis' team will be able to accurately track first pass yield data, lead time at the single product level, and real-time WIP tracking. IC

BEHIND THE BYLINE

Sean McCloskey is president and chief executive of Visiprise, Inc. in Atlanta.