Micros in manufacturing; personal computers and production control.
Personal computers and production control
The entire business environment is being overrun by the microcomputer revolution. The personal computer, in just a few years, has grown from a game-playing adolescent to an overpowering adult threatening even the minicomputer in capability. There's no question that micros are establishing themselves as ubiquitious are establishing themselves as ubiquitous from the executive suite on down to the factory floor.
The Yankee Group, Boston, MA, estimates the microcomputer market, including application software, will approach $4.5 billion this year. Approximately 63 percent of the market will cater to the business sectors, with 20 percent of these sales directed at manufacturing applications.
Importantly, the manufacturing planning-and-control software market (MPCS), which is approaching $35 million this year, will grow at a 25 percent compound annual rate.
MPCS is often thought to be the same as material requirements planning (MRP). This isn't so. MPCS includes the entire set of manufacturing control, process, and analysis applications (inventory management, bill of materials, production scheduling etc). MRP, although extremely important, is but one MPCS application.
When evaluating MPCS on a micro, the tendency is to dismiss it as a bargain-basement approach, with all the attendant shabby merchandise connotations. Yankee suggests, however, that equating price with quality in this case is bad business. Supermicros, combined with hard-disk storage, can handle the complexities of many MPCS requirements. And as raw processing power soars and prices continue to drop, these systems will render manual-scheduling methods obsolete.
In fact, because of the microcomputer revolution, all but the smallest manufacturers will be using micro-based MPCS systems to control the manufacturing floor by the end of the decade. Yankee projects that micro-based MPCS sales will exceed $160 million by then.
More than 20 suppliers now offer MPCS packages for supermicros. As in other manufacturing areas, competition for market share will be fierce, and effective competitors must meet one or both of the following requirements: provide links to IBM, Hewlett-Packard, and DEC equipment; support the CP/M and/or UNIX operating systems.
Currently, the market can be divided into the following three categories:
Low-cost systems. Suppliers like B-Squared and A Eric Eastman offer cost-effective packages that smaller manufacturers can use; however, the software can't be upgraded to multisite or local-area-network environments.
Specialty packages. Application specific systems, like Key System's Job-Shop MRP, are designed to address the needs of sophisticated users who are specific in their goals for MPCS implementation. Such products meet narrowly defined application requirements.
Modular MPCS. Most suppliers are offering a more comprehensive, modular approach to MPCS, where flexibility is important. These systems are powerful enough to move into the network or multisite environment, while remaining fast and responsive to the reporting and control needs of users. Twin Oaks, McCullough, RAIR, and Micro/MRP are a few modular-system suppliers.
A place for the commonplace
At the other end of the spectrum, Professors James H Macomber and Lawrence P Ettkin, both of the University of Tennessee, recently reported on the suitability of commonplace microcomputer software to handle sophisticated manufacturing production and control applications. Their presentation was part of the American Production & Inventory Control Society (Falls Church, VA) sponsored conference, "Microcomputers: tools for production and inventory management.'
The academics' experiment involved senior-level industrial-management students equipped with Apple II Plus micros and VisiCalo software. The purpose was to test the ability of factory planners to respond to a variety of materials- and labor-control situations without much background support.
Test participants were not given manuals or documentation for either the hardware or software. (Some people believe this is the acid test for user friendliness.) They were given, however, basic exposure to the operation of the hardware and software--enough so they could work out effective "what-if' situations in response to classic factory-planning problems. The results were encouraging.
All of the students, working in competing teams, were able to develop reasonable answers to a complicated production-planning problem using either linear-programming solutions or a transportation-algorithm solution (either of which would yield optimal production plans), or a trial-and-error solution (which might or might not yield optimal plans).
The researchers reached two conclusions: Manufacturing planning can be enhanced and well supported through this type of software; and personal computers are effective tools for solving a host of production-control problems, even when users have only a minimum of programming experience.
IBM PCs eye the plant floor
According to The Yankee Group, the Boston-based market research firm, IBM's XT/370 personal computer is clearly one of the most important industrial automation announcements of the past year, with vast implications for the manufacturing and engineering workplace. The key feature of the XT/370--a PC XT with three extra boards--is that one of its boards has several microprocessors containing the bulk of the 370 instruction set, which forms the basis of all IBM mainframes. For all intents and purposes, the XT/370 is the equivalent of a single-user 4300 mainframe.
Yankee believes the XT/370 provides a complete computer-hardware migration path from mainframes down to micros for manufacturing users. In addition the XT/370 is prime for pushing IBM's industrial-machine-programming language (AML) as a standard.
There hasn't been a significant advancement in manufacturing or machine-control programming languages since the development of APT in the late '50s. IBM (Boca Raton, FL) hopes that AML will be the basis for the next generation machine-control language.
PCs and robots
Programs for IBM's robots (the 7535, 7540, and 7545) are even being developed on vanilla PCs using a subset of AML, called AML/Entry version 3 (price is $3000). Loaded from the micro to the robot's processor, one PC can serve many robots since they don't depend on the computer for operational control. This foreshadows the personal computer's impact in the factory.
Medar Inc, Farmington Hills, MI, a control-systems builder and a value added remarketer of IBM robotic systems is using PCs for off-line programming. David Johnson, Medar's robotic-systems engineer explains, "Standard programming commands in AML are words you use everyday: Down, up, grasp, release etc. To program, we develop various subroutines to do whatever task needs to be performed--pick up a part, move from one point to another, place a part, and so on.'
To enable a programmer to position the robot, the work envelope is displayed on the PC's monitor, and it also shows where the robot is located. Starting with the robot at the home position, the programmer enters appropriate co-ordinates to move the end effector anywhere within the work envelope.
As soon as a program is written and compiled, it's downloaded to the robot's computer control via a standard RS-232C port. "Once the program is developed,' says Johnson, "the motion control needs to be interactively fine tuned.'
Most industrial robots today don't use personal computers for off-line programming. Typically a user employs the robot's on-board computer for both running and programming. Because of this, whenever a robot is being programmed, it's out of commission. The advantage of using a PC, when it can be employed, is that the robot can work on a job while a program for another application is developed.
"Off-line programming,' remarks Johnson, "allows a robot to be used more efficiently. Moreover, it's faster and more convenient. It might take, for example, seven days to develop a sophisticated application program, and all during that time the robot isn't running. Using the PC, it might take just four days to write the same program, and just one day working with the robot to refine it.
"I believe off-line programming is the preferred way with robots; however, I don't know if PCs will be used universally for this purpose. One company, for instance, uses their own special controller for off-line programming. The problem is that there aren't any industry standards; everybody uses a different language and a different protocol. But, when you move into CIM, off-line programming is definitely the way of the future.'
SPC and the PC
In a different application, GSE Inc, Farmington Hills, MI, is using IBM PCs for generating statistical process-control reports from data gathered with the firm's torque-monitoring systems. According to VP Harold Munn, "The purpose of the reporting system is to enable customers to determine if their processes are capable of meeting engineering specs, and then maintaining control of the processes.'
Developed for use with the firm's Model 238 and Model 575 microprocessor-based monitors, torque data is collected by the monitor and then transferred to the PC. There are three methods for transferring information: Directly from the monitor to the micro through a standard RS-232C port, from the monitor to a cassette tape and then to the PC, or from the monitor to a solid-state memory device and then to the PC.
Once in the PC, the data is reformatted so a special program can perform the necessary calculations and produce the statistical reports. Graphics that can be generated include standard-deviation control charts; control charts fr averages with upper and lower control limits, and control charts for range; histograms (or frequency distribution charts); and tolerance charts. A process-capability report also can be created. All can be displayed and printed.
"The SPC software,' states Munn, "creates a powerful tool for manufacturing process management and control. With the growing use of automated systems, we expect to see micros applied in many areas, taking over more control and reporting functions.'
Chrysler cashes in on programmable welding controls
As US automakers continue retooling old plants and opening new ones, they are equipping the facilities with state-of-the-art machinery and control systems to maximize efficiency, flexibility, and quality. A new Chrysler assembly plant (Windsor, Ont, Canada) is applying such equipment by using microprocessor-based resistance-welding controls manufactured by Medar Inc, Farmington Hills, MI.
Over 250 of the controls are on a spot-welding line for producing the front-wheel-drive Plymouth Voyager, Dodge Caravan, and Mini Ram Van. The automaker also has installed nearly 200 more of the controls at its Sterling Heights, MI, assembly plant, where the intermediate-size Chrysler Le Baron and Dodge Lancer passenger cars are manufactured.
The welding controls can be programmed for up to 32 weld sequences, with 99 weld functions per sequence. Using a key pad, together with a 32-character LCD display, an operator follows prompting messages that guide him through the necessary programming entries. The display also indicates current weld status and identifies weld problems. Because of the programmability, Chrysler can produce additional car models at each plant.
Flexibility of the weld controllers enables handling a variety of welding jobs and material types, including galvanized steel and zinc-coated metals. The units can be used with one- or three-phase, single- or multiple-point welders.
Also, a heat-control system--consisting of zero-crossing detection circuitry, automatic voltage compensation ( 20 percent), linear heat control, and power-factor correction--virtually does away with misfires. To eliminate drift, a quartz crystal is used as a reference for all timer and counter functions.
Two RS 422/232-compatible ports allow interfacing with a host computer or sending files to a printer for hard copies of sequence-, setup-, stepper-, and weld-performance-summary data.
A final feature is the array of fault indicators, which reduce downtime by helping maintenance personnel quickly pinpoint problems. Users can define these indicators and specify fault significance, e.g., major or minor.
Microfinishing straightens up
In the past, if you delivered an egg-shaped part to a microfinisher for finishing, you received a polished, egg-shaped part in return. Now, using the new GBQ-2000 process, it is possible to correct the geometry of the part during microfinishing, correcting for lobing as much as 60 to 75 percent. The new process also eliminates oil-hole guttering washout, which is generally the result of conventional polishing methods, according to Industrial Metal Products Corp. Lansing, MI.
IMPCO says that GBQ stands for Generating Bearing Quality, and, indeed, the process could set new standards for polishing crankshafts and other drive-train components. Historically, engine crankshafts called for 16 or better microinches--industry state of the art. Now, using the GBQ-2000 process, these workpieces can be polished down to 5 micro or better. Steel parts, which do not have nodules, can be polished to micros of 1 to 3.
Ferrite caps found in nodular crank-shafts have always been a major problem. These nodules are abrasive particles that protrude above the surface of the bearing. If they are not removed, they begin to cause damage to the crankshaft within the first few hours of operation. The new process removes up to 90 percent of the caps, and this fact alone would make the process very important.
New testing equipment used to verify the improved performance includes a computerized surface roughness and waviness analyzer that prints out up to 25 different inspection results. Such equipment proves that the new process can reduce the R(A) (roughness average) from 17 (using previous processes) down to 4, using GBQ-2000.
The new process also reduces the skew value (RSK) from 0.1 down to -4.3. The lower value indicates a more-suitable bearing surface. Furthermore, there is a dramatic improvement in bearing area (Tp) compared to previous methods.
Grinding to millionths
With a new cylindrical grinder, Jewett Machine Mfg Co, Richmond, VA, is able to grind round parts to millionths of an inch, both for extensive contract work and parts for their grinding machine, an end-mill grinder.
The machine, a Karstens Series K10-11, features electronic infeed and advanced plunge and traverse grinding cycles. Its universal design makes small-lot grinding economical as well. The wheelhead swivels for conversion from ID to OD face grinding in 30 sec.
The accuracy of the headstock and the use of high-precision bearing make possible the grinding of heavy workpieces as easily as small ones.
Buick City's pulsing pistons
New piston-turning technology at GM's Buick City, Flint, MI, complex hopes to outmode present piston production and improve engine performance. The first of a new twin-line dedicated transfer system, it will produce completely finished aluminum pistons from foundry-turned castings using only two multioperation machines.
Built by LaSalle Machine Tool, Troy, MI, the system is expected to improve piston quality significantly, reduce manufacturing overhead, and reduce floor space use, replacing 28 machines currently in use.
The system consists of two, 22-station lines of metal-removal equipment, each feeding a common gaging/classifying unit. It can process a family of different pistons at total rates of 1500/hr, including two different parts simultaneously per line.
Diamond tooling and turning technology enables the system to machine complex pistons to tolerances previously unattainable, and with only three operators attending, it can achieve production uptime levels of 95 percent.
The LaSalle piston-finishing machine completes operations previously done by three machines. In one 30-sec clamping, it turns and chamfers the ring grooves, turns the complex skirt shape, bores the wrist-pin bore, and holds dimensional relationships rigidly uniform and square with each other. Typical tolerances are: bore and groove parallelism to dome within 0.002, wrist-pin bore size within 0.0003, and skirt dia complex shape within 0.0006.
When fully operational, only one operator is required to load both lines of the system:
Three-spindle orbital headforming machine has a programmable controller for accurate point-to-point positioning and automatic X-Y indexing. It was developed by Taumel Assembly Systems, Patterson, NY, for a manufacturer of locks and related hardware.
Called the System 323/CNC-80-6, it has three single-spindle modular heading units, each with different quick-change form tools to flare out solid studs and rectangular tabs, secure bushings, or caulk/stem shoulder pins in lock plates.
It is used to assemble six types of lock plates with varying hole patterns; the largest is 11 8. An automatic X-Y table with 20 12 travel is driven by stepless DC servomotors. The CNC controller can recall existing programs instantly or be reprogrammed in minutes, greatly reducing changeover time.
The system replaces several noisy stakers and air presses that could fasten only one stud, pin, or bushing at a time. This required multiple handling by many operators and 100 percent inspection. Now, only one operator positions components in a fixture nest while the previously loaded lock plate is being assembled in the CNC machine.
New surface treatment
With a new surface treatment process developed by Lucas Industries in England and called Nitrotec, plain carbon, medium carbon, and low-alloy steel parts can be given increased yield strength, high corrosion resistance, increased lubricity, and a visually attractive, matteblack surface finish. It is particularly suited for applications in the automotive, aerospace, appliance, recreational-equipment, and agricultural-equipment industries.
US licensee Ipsen Commercial Heat Treating, division of Ipsen Industries, an Alco Standard company, feels the increased yield strength will make possible the use of smaller and lighter parts without the need for subsequent painting or zinc, chromium, or cadmium plating. Because the process is low temperature, it offers improved wear and fatigure properties with very low distortion compared to nitrocarburizing, carbonitriding, carburizing, and induction hardening.
Urethane-faced tools outlast epoxy
At Excel Pattern Works, Los Angeles, CA, a significant extension in tool life for molds used to produce cruise-missile fuel tanks was made by using new urethane wear-resistant surface coating. Called RP-6414 from Ren Plastics, E Lansing, MI, it helped urethane-faced tools last ten times longer than standard epoxy-faced tools.
Using polyurethane as a surface facing was a logical step because of the material's resistance to blown sand, its low cost compared to metal, and its ease of use. It is not affected by humidity and dryness, doesn't oxidize, and has shock resistance to prevent tool faces from chipping when knocked or dropped.
Excel's mold-building process starts with carved mahogany mold halves. A paste-wax coating serves as a release agent, followed by a brush coat of petroleum jelly. The RP-6414 urethane is brushed on in thicknesses from 0.020 to 0.060. The joined mold halves are supported on plastic strips in an aluminum container for the pouring two hours later of RP-1710 epoxy to complete the mold. While the epoxy is still wet, glass fibers are spread over the tools and tamped with a brush to add strength.
A new high-energy metalworking service at Battelle Memorial Institute assists companies in fabricating products, selecting and forming materials, and establishing product specs using explosive welding and compaction. For 20 years, these technologies used explosives to transmit high-energy impulses that bond metals, compact metal and ceramic powders, or form metal shapes. They have been used primarily to produce diamonds and flat-plate metal cladding.
"Explosive metalworking technologies produce true metallurgical bonds that will be needed in producing parts for future electronics, chemical, energy, transportation, and military applications,' says Battelle's Dr Narain Lalwaney, who oversees the service. "The technology is available. Our purpose is to help organizations apply it.'
They will be concentrating initially on bonding and dynamic compaction of metals and ceramics. In forming, they will be offering help in sizing, punching, autofrettage; the redistribution of stresses; and shock hardening. The service is aimed at helping both small and large organizations, for both short-term projects and long-term technology development.
Casting the die for quick die changing
Human-powered flight enthusiast Dr Paul MacCready is fond of explaining how he collected the $200,000 Kremer Prize for the first successful human-powered flight across the English Channel in 1979. While others tried to perfect aerodynamics using high technology, or sought highly skilled pilots whose flight savvy might overcome design limitations, MacCready walked away with the money using a simple--almost clumsy-- design; his human motor was a professional cyclist. Because he defined the problem differently, he won the prize.
The story is worth remembering when setting out to solve problems on the shop floor. For example, in American industry, discussions on the subject of increasing productivity usually center on high technology solutions, often at the expense of simple ones.
One major drain on productivity that lends itself to a simple solution, especially in small- to medium-sized metal-stamping plants, is die changing. It's a necessary evil that can easily consume hours that are then added to the accounting burden carried by each part produced.
A typical metal-stamping press might cycle at 100 strokes/min (6000 pcs/hr). So a typical die change, consuming four to six hours, is an enormous drag on profits. Production managers frequently overcome this problem by lengthening the lot run for an ordered part, betting that the customer will reorder in the near future. Here, the manufacturer accepts inventory costs until final delivery is made, and that can get expensive.
Bob Proost, an applications engineer for Vlier Engineering's Dayton, OH, office (a manufacturer of hydraulic clamping systems) notes, "One of our clients counted $2 million in finished product that they were carrying as inventory in just one division. Instead of earning a good return in a high-interest period, they were losing use of that money. That revelation is responsible for the growing interest in just-in-time (JIT) inventory control--a simplified approach to minimizing inventories, which was developed in America before WW II, but exploited and perfected in postwar Japan's major industries.'
When pressworking, JIT relies on quick die changes to produce parts just as inventories are depleted. To make it work economically, the Japanese have cut average time needed to change dies from hours to minutes by standardizing such things as die-shoe thickness and mounting plates on a family of presses. Quick-acting hydraulic clamps then are used to hold whatever die is selected for the press.
With a die-set clamping system, composed of off-the-shelf power workholding components to hold the top and bottom dies to the press's ram and bolster plate, a relatively small investment of from $5000 to $10,000 can cut die-changing time to only a few minutes. In addition to the clamps, the only other elements are an air/hydraulic power source to generate pressure, small plumbing lines that carry the hydraulic oil from an intensifier to the clamp, and sequencing valves (if necessary).
A die-set clamping system will pay for itself in three to six months, claims Proost, who has been installing such systems at companies such as Chrysler, Bendix, Westinghouse, A-C Sparkplug, and more than two dozen other shops.
Die-set clamping principles are simple. The problem is the same as when conventional die retention systems are used: Devices are selected that hold the dies firmly despite pressures and movement of the press ram. Press speed, stroke length, and weight of the ram and dies all must be considered. Vlier has reduced such considerations to a simple formula that takes less than a minute to work through, and is detailed in a free four-page bulletin called "Stop wasting expensive press time changing dies.'
Case in point
For complete exploitation of die-set clamping's benefits, press plates and shoe thicknesses are generally standardized. The result is virtually any die in the shop can be dropped into any press.
That's the aim of McCord Gasket Corp, (a subsidiary of Ex-Cell-O Corp), Wyandotte, MI. Some months ago, reports Andrew R Skrycki, McCord's master mechanic, the company launched a die-set clamping program after determining that die changeovers were draining productivity. The results have Skrycki and others at the company enthusiastic. "We were spending several hours doing die changes and setup,' he points out. "Many of our dies are progressive, involving three to eight operations. We're now averaging only 25 min per change with the die-set clamping system.'
Best of all, each die change is accomplished by only one person, the press operator, who removes the die at the end of a run, exchanges it in the die room for a new die, then returns to the press to install it. While that adds minutes to the die-changing operation, it's more cost-effective for McCord than the Japanese method where a die-change team swarms around a press, completing their mission almost immediately.
The McCord plant contains Minster, Niagara, Bliss, and other OBI presses. The company is looking to standardize, so eventually any operator in any of the plant's departments can work on any press with any die--the ultimate in efficient use of men and machines.
Vlier's Proost, noting this example, emphasizes that die-set clamping's benefits are so extensive it's as much a management concept as a toolroom technique. "One company increased productivity on some of their presses by 15 percent using this concept,' he reports. "It only takes a handful of presses producing 15 percent more to eliminate the need for buying a new press. Not only can the costs of financing unnecessary inventories at double-digit interest rates be minimized, but the opportunity costs associated with buying new equipment are obviated.'
JIT: No room for error; CMMs help
It's estimated that the cost of work-in process inventory is second only to the cost of direct labor in manufacturing facilities. Consequently, manufacturers are turning to increased levels of automation to reduce the labor costs in the equation, and to just-in-time (JIT) production systems to cut inventories.
JIT production systems require that workpieces are available for machining and parts for assembly only when needed. Generally, discussions about JIT operations center around flexible machine tools, automated material-handling equipment, robots for assembly, and the like; however, according to engineers at DEA, Livonia, MI, coordinate measuring machines are critical in any JIT production system, and so shouldn't be overlooked. The reason is there's no extra inventory available. Parts must be made right the first time.
A spokesman from DEA says, "CMMs can spot the onset of problems in manufacturing operations before they become trouble, i.e., catch deviations before they go out of tolerance. This permits parts to move through the system as they are intended.' By including CMM stations in a manufacturing system, dimensional characteristics can be checked to ensure parts are machined correctly and to permit required adjustments to specifications are maintained.
A new powder-metal (PM) forming system can manufacture ball-bearing assemblies with bores eccentric to the raceways. The patented forming process densifies and strengthens the material in the critical raceway area while maintaining low production costs. The formed rings are heat treated and can be ground to meet the quality required for each application.
The race-forming process allows the PM inner rings to be made with offset bores and a variety of positioning features such as keyways, integral keys, flats, and tapers. The process eliminates (1) the need for expensive broaching of keyways or flats and (2) the use of offset shafts or inserts to achieve eccentric motions. Fafnir Bearing Div, Textron Inc, New Britain, CT.
Digitized cost estimating
You can achieve fast, accurate estimates with "sonic digitizing' using the MT Costimator(TM) system. The first step using the system is to enter standard machining data such as machining methods, speeds, feeds, tooling, and material. Once these facts are stored in memory, the operator can simply trace over a blueprint with a pen-like device to find length of cuts, the depth of a bored hole, the cutter path of a milled part, or the length of a bar to be turned.
As the user emulates the cutter path with the sonic pen, the digitizer traps the X and Y coordinates. This information goes to the microcomputer and is converted to metalcutting time and machining cost. The system is said to provide up to 75 percent reduction in the time required to produce a conventional estimate. Payback (ROI) periods of 20 to 40 weeks are to be expected, according to Manufacturers Technologies, Springfield, MA.
Centerdrive lathes speed handling
Automation advances in Centerdrive lathes bring part-handling efficiencies and CNC four-axis flexibility to these machines that enable you to simultaneously machine both ends of a workpiece. The Model CD-500 by Seneca Falls Machine Co, Seneca Falls, NY, is integrated with a special heavy-duty gantry loader.
The machine features touch-off tool compensation and a unique high-pressure pulsating coolant system that provides good chip control and extended tool life. It's offered in 25-hp and 50-hp capacities, and even larger units are being developed.
Centerdrive headstocks are available with spindle bores from 6 to 15 and up. Operations possible using independently indexing turrets at each end include short OD turning and threading of ends, facing, chamfering, ID boring and coboring, and ID threading.
A typical operating sequence begins with the rough part being picked up by one loader arm while the previously machined part is discharged at the same point by a second arm. As the machining cycle ends, the spindle stops, the tool slides to the rear, and a part-support rail indexes into position beneath the spindle. An injector/extractor head moves into the bore of the part, expands, and draws the part out of the spindle onto the support rail. As the extractor head collapses and clears the part, the loader arm moves down and picks up the machined part while a second arm moves down and places the new part on the support rail.
Photo: New GBQ-2000 microfinishing system at work, getting surfaces down to a few micros.
Photo: Tough test equipment follows through with fast, detailed inspection.
Photo: A die shoe is held to the bolster plate utilizing a T nut, stud, and hollow hydraulic cylinder.
Photo: A strap clamp is used with a hollow cylinder to move the cylinder away from the die.
|Printer friendly Cite/link Email Feedback|
|Publication:||Tooling & Production|
|Date:||Sep 1, 1984|
|Previous Article:||Optical quality FMS.|
|Next Article:||Flexible assembly pays off.|