August 4, 2009
Addison Machine Engineering, a manufacturer of tube and pipe mills and tooling, found itself struggling to retain new customers and satisfy existing ones. Although the company was long on experience and embraced current technologies, the company had difficulty in expanding its market share.
"We used to lose ninety percent of our new customers," said Gerald W. "Jay" Brunken Jr., vice president of operations. "That locked us into a no-growth mode."
A customer's audit in 2003 provided an outsider's perspective. "They focused on our machining, which we thought was pretty good," said Robert Clyde, vice president for quality. The audit used green, yellow, and red to indicate Addison's capabilities. "Most of what we got was yellow shading over into red," Clyde recalled. "We didn't do very well."
Management knew it had to put the company through a major overhaul. It started by focusing on its machining operations, and for good reason. Machining roll tooling—new rolls and regrinds—generates 80 percent of the company's sales.
A primary challenge concerned at-the-machine (A-T-M) programming. Some machinists were unable or unwilling to program a machine while a job was running, leading to unnecessary downtime. In addition, errors, scrap, and rework were common because the working surfaces of many rolls are hard-to-program, geometrically complex curves such as hyperbolas.
A recommendation to use Edgecam to computerize machine tool programming and move it offline brought in MFGtech, the local Edgecam reseller.
MFGtech developed about a dozen program command integration (PCI) macros, which have almost completely automated many programming sequences. They allow the programmers to select the roll geometry, turning centers and cutters, and workholding setups. The macros automatically regenerate the geometry in Edgecam; generate toolpaths; and turn them into machine code.
By 2005 nearly all A-T-M programming was gone, but productivity problems lingered. "The bottleneck was poor material flows and low productivity on both a machine tool and individual worker basis," Brunken said MFGtech sent in Ben Black. His approach was a mix of 5S, lean, value-stream analysis, and cellular manufacturing.
The first, 5S—sort, set in order, shine, standardize, and sustain—covers basic housekeeping and is the starting point for kaizen, Japanese for continuous improvement; lean manufacturing slashed work-in-process and raw material inventories, the latter a result of ordering steel only when needed; value-stream analysis mapped every manufacturing step, weeded out every activity that did not add value, and helped pinpoint root causes of production bottlenecks; and cellular manufacturing led to organizing machine tools into five production cells by workpiece size. Finally, a wireless system integrated machinery with managerial functions such as time keeping, production processing, and routing.
Results include a 25 percent increase in on-time delivery; a 75 percent reduction in setup times; a 50 percent reduction in secondary operations; a 60 percent reduction in scrap; and a $300,000 inventory savings (in the first year alone) by limiting steel purchases. These steps also eliminated several work-handling steps and reduced labor-hours with simplified routings.
The wireless installation eliminated cable hassles when AME relocated key production machines into cells.
The biggest cost savings was the planned construction of a new, larger facility. The company was running out of room in its current 25,000-sqare-foot facility. MFGtech showed the company how to develop workcells, improve workflows, and get inventories out of the way.
"It would have cost us $1.8 million to $2.4 million to build a new 40,000-square-foot facility," Brunken said.
Reducing costs is one thing; increasing revenue is quite another. Management was pleased to find that these changes had improved the company's potential to increase its revenue by expanding its market share, as measured by the number of new customers it now retains.
"We keep ninety percent of our new customers," Brunken said.