Getting it under control

Modern Machine Shop, July, 1991 by Tom Beard

Getting It Under Control

Computer-assisted shop control smooths out this job shop's internal workings, gains the respect of their customers, and brings additional dividends to the bottom line. Surface Engineering is hardly your average shop. Their specialty is high-precision grinding of small round parts--just the sort of thing you'd expect to find in computer disk drives. No wonder, then, that this San Jose, California, shop does a substantial portion of their business with most of the major computer and peripherals manufacturers also located there in the heart of the Silicon Valley.

But once you get beyond the state-of-the-art grinding and measuring equipment necessary to hold sub-tenth tolerances, this shop is pretty much like any other. They have to quote quickly and accurately. They have to manage their work flow. They have to make a profit.

Complicating matters, however, is an industry-wide trend in the computer business to employ just-in-time (JIT) manufacturing operations. For the supplier, that means a typical order cycle, from quote to delivery, that once could take several months now must happen in three or four weeks. Customers simply demand that kind of performance. And everything has to be right. No late shipments. No rejects. No excuses.

To perform to these standards, a supplier must have its shop floor and business mechanics well under control. Surface Engineering's president, Dick Peattie, saw it all coming five years ago and began taking aggressive steps to get a better handle on his business. One of the first actions was to computerize the company's quoting, scheduling, routing and financial activities with a software package that integrates all these functions into a single database. Dick Peattie was betting that better management of the business would begin with better management of information.

Through The Shop

The software system Surface Engineering uses was developed by Job Boss Software, Inc. (Minneapolis, Minnesota). It is designed for job shop-type operations, supporting estimating, scheduling, job tracking, direct costing and billing. Many shops run such a system on a single personal computer, but at Surface, they have a network that provides all the key system users ready access to the shop data via PCs on their desks. That access is most important to production manager Steve Pinick, who is charged with loading and managing the shop, and to Mr. Peattie, who is focused on sustaining the competitive stature and financial well-being of the company.

Following a typical order through the system helps explain how Surface captures information, and then puts it to use. To quote a job, Peattie and Pinick discuss the part and come up with a general process plan, including anticipated cycle times. When Mr. Pinick enters this data, the system generates all internal costs based on standard shop and overhead rates they apply to each job. If delivery is a concern, he can look at the scheduled loading of each workcenter to make sure a machine is not over-committed.

Next, office manager John Gilmore will access the job file. He will check the cost of all materials and services bought on the outside, making sure availability is within their needs. Once he enters these outside costs into the system, it is ready to generate a formal quote. They use standard rates for labor and markup, but these may be modified if, for example, a rush order will disrupt the shop or stringent quality standards will require the attention of their most skilled, and expensive, people. Two- or three-day turnaround on the quote is typical.

When the quote becomes an order, the router is generated directly from the process data already created. Steve Pinick then details the plan, adding more specific instructions so the job requirements will be clear to each operator in the process chain. At the same time, he enters the job into the system's production schedule. Because of their customers' JIT programs, Surface schedules tightly around a desired delivery date. The system aids this function by beginning with the date, and then scheduling backward through the process steps listed on the router. Thus, run dates are automatically generated for each step, based on the specific run quantity and cycle times on the router and some averaged assumptions on setup times at the various workcenters. Production yields are also figured backwards. That is, they may assume some scrap at each station (due to setup, for example) and then build those factors into the starting quantity and internal cost summary.

The obvious advantage of this method of infinite capacity backward scheduling (see Infinite and Finite Scheduling) is that it addresses the very practical consideration of making sure the early process steps are done in time to keep the job on track. The disadvantage is that it is quite possible to over-commit a workcenter for any given time period. The system does have the ability to generate a schedule based first on capacity restraints, rather than completion date. The point of this method, called finite capacity scheduling, is to more efficiently utilize the shop's workcenters.