Engineering News 7508Engineering News 7508
September 22, 1997

Coatings keep pace with industry demands
Newton, MA--When it comes to the machine tool industry, two trends stand out: Faster spindle speeds, and a goal of dry or semi-dry machining. The technology allowing these advances? Continuous improvements in coating chemistries, as well as better processes for applying these new coatings.
For example, as a general-purpose workhorse, TiN (titanium nitride) covers a wide area of applications in the machining of ferrous materials. TiN adds an as-deposited hardness of approximately 80 HRC and withstands service temperatures of up to 1,100F (593C). Bohler Edelstahl GmbH, an international steel company headquartered in Kapfenberg, Austria, advises customers that properly applying TiN coatings can prolong tool life by a factor of 3 to 8 times.
Should wear problems persist after applying TiN, Bohler advises upgrading to a higher-performance carbon-based or TiCN (titanium carbon nitride) coating. TiCN coatings boost hardness to 90 HRC, and manufacturers recommend them for punching and forming tools where excessive stresses are encountered.
"After titanium nitride, a series of coatings has been introduced that performs much better than TiN in niche applications," says Fred Teeter of Balzers Tool Coating's Wear Protection Division, North Tonawanda, NY. Examples include Balzers Balinit(R) Futura and X.treme brands of coatings. Both are based on TiAlN (titanium aluminum nitride)--an extremely heat- and oxidation-resistant alloy.
Futura is meant for carbide, high-speed steel, and cermet tooling; X.treme suits carbide milling of hardened workpiece materials. Though both are TiAlN, Futura represents a multilayer coating of varying chemistries while X.treme is formed from a uniform single layer.
By slightly altering the chemistry within a coating--that is, by changing the stoichiometric ratio--a manufacturer can produce a whole family of TiAlN compositions, each one offering slightly different densities, hardnesses, and other properties. Additionally, the deposition process used to form the coating determines crystalline structure, porosity, and other features affecting coating characteristics.
"Ten years ago," Teeter says, "the only PVD (physical vapor deposition) coating available was TiN. Now, there are at least a dozen. And when you start to use them in multilayer combinations, you can get even more." The trend, he predicts, will be an increasingly wider range of coatings and coating combinations--each one optimized for a certain application.
Lubricity lowers cost. Many economic studies show that average cost of the cutting tool as a percent of total cost of the machining operation is about 4 to 6%. The metal-working fluid cost as a percent of the total machining operation--including purchasing, handling, and waste removal--represents up to 16%. In effect, coolant cost can be 4x the tool cost. The economics of cooling, therefore, is driving the use of coatings within the machine tool industry.
Health concerns are another driver. Because some metal-working coolants and lubricants contain known carcinogenic materials, many companies are looking to lower exposure levels, usually by cutting back on the coolant level. One way to compensate for less coolant: increased use of coatings for semi-dry or even dry machining.
Example? HANIT 8, a TiAlN-based coating with doping material that increases corrosion stability during cutting. Developed for cemented carbide tools by Hauzer Coating Centrum, Vehlo, The Netherlands, HANIT 8 accommodates an additional metal-containing carbon hydrogen coating with chromium or titanium as the metal component. "This type of coating," claims Hauzer Coating Director Werner Fleischer, "significantly reduces tool friction for semi-dry cutting of Al-Si alloys."
Non-tooling applications. As with machine tools, precision components are subject to increasingly more difficult operating conditions. For example, strict emissions regulations have meant fuel injectors designed for higher pressures and temperatures and having ever tighter tolerances. The danger for the design engineer becomes one of developing a component that will pass emission tests but fail life tests due to wear.
Coatings, again, can provide the solution. Consider the ideal screw compressor. Unlike oil-injected compressors that operate without a synchromesh gear, or dry-running compressors with a synchromesh gear, the ideal compressor would combine the best features of both: oil-free air and no expensive synchromesh gear.
On the way to this ideal design, industry is turning to water-injected screw compressors that utilize a synchromesh gear. Advantages of this approach, compared with oil-injected designs, include oil-free air and higher mechanical efficiency due to lower viscosity. Benefits over dry-running compressors include a higher pressure ratio, smaller sealing clearance, and lower noise.
Next step to the ideal would be a wear-resistant coating for water-injected compressors without a synchromesh gear. Such compressors, tested with Balzers Balinet(R)C Tungsten Carbide/Carbon (WC/C) coating, indicate this is possible. After 1,000 hours, compressor screws showed only 3% surface wear. Projected service life: more than 20,000 hours.
Similarly, Hauzer's HAN-IT 10 CrN-based coating has demonstrated effective wear and corrosion resistance, as well as low friction, on automotive connecting rods and piston rings. Other precision components such as gears, tappets, and camshafts face the same demands with regard to ever-higher performance criteria. As with machine tools, their ability to meet these requirements rests with a thin layer of today's--and tomorrow's--many new coatings.
Properties Comparison
Trade Name | Balinit(R) FUTURA (TiAlN) | Balinit(R) X.TREME (TiAlN) | Balinit(R) B (TiCN) | Balinit(R)A (TiN) |
---|---|---|---|---|
Microhardness (HV 0.05) | 3,000 | 3,500 | 3,000 | 2,300 |
Maximum Working Temperature | 800C/1,470F | 800C/1,470F | 400C/750F | 600C/1,110F |
Key Characteristics | Multi-layer | Single-layer | Multi-layer | Single-layer |
Applications | Carbide and high-speed steel, high-speed and dry machining | Carbide endmills, machining of hardened steels (Rc 45-65) | Full-range, high-performance with coolant | Full-range, general purpose |
Tough skin permits portable, reduced-glare lighting
Poisant, France--Drive by any nighttime construction site and you'll know how highly contrasted spotlight illumination can be. Now Airstar offers a portable, self-inflating "illumination balloon" to provide reduced-glare lighting for night operations.
Key to the Sirocco balloon is its triple-layer, high-tensile-strength polyester skin, which encloses a 1- or 2-kW halogen light. Benoit Beylier, export manager for Airstar, notes the difficulty of finding the right skin material: "There was nothing on the market good enough at combining the proper light-transmission, fire-resistance, and temperature characteristics." Porcher-Marine (Lyon, France) was able to supply a custom polyester skin that met all requirements.
In addition to strength, the polyester balloon skin has to have the correct light transmission and airflow properties to permit adequate cooling. |
Mounted on an adjustable stand 2.5 to 5m (8.2 to 16.4 ft) high, the lamps light 11,000 to 15,000 ft2. The upper balloon surface is aluminum plated to serve as a reflector. The white, translucent lower skin widens the apparent light source for more uniform illumination, producing faint shadows while cutting back glare. Within the balloon a protective grid protects the lamp .
Sirocco attaches to a tube stand or clips onto a worksite machine. An internal compressor inflates it within 45 sec. "The blower had to be small, efficient, and strong enough to maintain pressure, and shape, in a high wind," notes Beylier. Sirocco's 43 x 32-inch oblate shape gives it stability in winds up to 60 mph. A pc-board-mounted pressure sensor controls the compressor to compensate for minor skin tears.
Consumer electronics gets on the ball
Lowell, MA--The Spaceball arrived in the early 1990s and proposed to do-in the infamous dial box as a multiaxis CAD controller. Innovative optoelectronic sensor-driven input let users navigate precisely through 3-D geometry while keeping their eyes on the screen and their hand on the ball. The only problem: The Spaceball cost thousands of dollars.
Now, Spacetec IMC Corp., inheritor of the Spaceball, is producing the 6D SpaceOrb 360 game controller based on the CAD input device. Lower-end versions, such as the $700 Spaceball 3003, have come to market, but the playtime model is intended for gamers, who typically shy away from multi-hundred dollar price tags. Through design innovations and high production runs, Spacetec managed to get the SpaceOrb 360 down to under $100.
"We knew we could transfer our 6D controller technology from our Spaceball engineering products to create a game controller," said John Hilton, Spacetec's chief technology officer. "Our challenges were to design a device that would cost under $100, retain the feel of the Spaceball, accommodate a wide array of hand sizes, and stand up to the frequent use and abuse during life-or-death game play."
The first step toward reducing costs was to develop an integrated circuit that would directly interface with the PowerSensor optoelectronic sensors and provide a simple digital interface to a standard microcontroller. Spacetec called on its Chandler, AZ-based Signal Processing Group, which was able to reduce the control circuitry to a single IC with only two interface signals. This device not only reduced electronics cost but maintained 10-bit resolution while reducing space and power needs.
Adapting the PowerSensor presented a more complex mechanical design problem. Six sensor mechanisms had to be arrayed in pairs along the x, y, and z axes. In addition, a dozen electronic components had to fit around them within the small volume of the ball. Hilton responded with a design, dubbed the "origami PCB" that folds the pc-board circuitry with the infrared detectors into a cube that fits inside the ball.
Due to the miniature design and dense componentry, issues of clearance and tolerances were critical. Designers used Pro/ENGINEER from Parametric Technology Corp., Waltham, MA, to construct an assembly model they could rotate to visually examine from all viewpoints.
Material delivers tough case
Eersel, The Netherlands--When one of SPI BV's top distributors, Faes BV (Bladel, Netherlands), needed static-dissipative cases, the firm turned to RTP Co. (Winona, MN) and its Benelux distributor, Distrupol BV, for a solution.
For transporting and storing sensitive electronics, the company needed cases made of a tough material with a surface resistivity of 107{OMEGA}. RTP 100 Series polypropylene copolymer met that requirement. The compound exhibits an un-notched impact strength of 20 ft-lbs/inch (1,068 J/m) at 1/8 inch and a tensile modulus of 200,000 psi (14,000 kg/cm2).
SPI BV, Distrupol BV, and RTP Co. worked closely to develop the material. Distrupol BV facilitated the formulation, testing, and specification of the material for the SPI project. The suitcases are available in eight sizes with exteriors that can be imprinted, depending on the application.
Diamond-impregnated coating improves wear resistance
by Charles J. Murray, Senior Regional Editor
DeKalb, IL--On some components, frequent replacement is a given. If the parts aren't expensive, and if the cost of downtime isn't high, then users learn to live with it.
But not all parts can be easily replaced. To help improve the wear resistance of those parts, engineers from The Armoloy Corp. have developed a coating that protects costly metal parts from wear, abrasion, and corrosive stress. The new coating combines extraordinary hardness with lubricity to increase wear resistance by as much as 10 times in some cases.
Key to the new technology is the use of tiny spherical diamond particles. The diamond particles, which measure no more than 5 angstroms in diameter, are added to an existing coating known as Armoloy TDC Chromium. The resulting product--Armoloy-XADC--has exhibited Rockwell hardness measurements in excess of 90 Rc.
Early users of the technology have applied it to cavity molds and bearings, but Armaloy engineers say it could serve in a wide variety of highly abrasive applications. In those applications, it serves as a harder, more lubricious, and more costly alternative to hard chrome or electroless nickel.
"The diamond impregnation enabled it to hold up in our applications, where other coatings were too soft," notes Al Horn, tool room manager for Hoffer Plastics Corp., South Elgin, IL. Hoffer employed the coating on steel cavity molds used to make aerosol valves. Prior to using the coating, the $5,000 molds were wearing out at a rate of approximately one per month. XADC coating extended the life of the company's molds to about nine months, Horn says.
The frictional coefficient of XADC against XADC can be as low as 0.09. |
Armoloy engineers, who have extensively tested the coating, believe that its lubricity is as important to its performance as its hardness. The XADC application process, they say, provides a nodular texture that creates a ball-bearing-like surface. The point-to-point surface contact of it against other materials lends itself to reduced friction, increased lubricant retention, and lower operating temperatures for parts. The frictional coefficient of XADC against XADC can be as low as 0.09. "The lubricity of this material reduces friction down to the level of PTFE," notes Armoloy President Jerry Bejbl.
Early applications for the new coating include food-processing and gasoline-handling equipment. In both cases, the material's resistance to corrosion and wear are critical. Ultimately, the company expects to see the material coat gears, tools, dies, drill bits, and hundreds of other metal machinery components. "This is a more costly coating," Bejbl says. "But if you need wear resistance and durability, XADC will give it to you."
Low-cost eye tracking offers myriad 'look-and-click' uses
Oldham, England--Eye-movement tracking is nothing new. But such a system costing well under several thousand dollars is. Originally developed by Ferranti Technologies Limited for fighter pilots, patent-pending technology from Vision Control Systems (VCS) is not the result of a major breakthrough, according to business manager Richard Gay, but rather of several technologies coming together. The company has formed a business unit, Vision Control Systems, to seek licensing partners for mass-production applications.
Previous eye trackers were costly, says Gay, and thus used mostly for physiological research. They track by recording a series of images of the eye and then processing them to project pupil path and determine its position. The number crunching for this method requires extensive, and costly, computer and software support. The simplicity of VCS does not require eye-image processing. A pupil-center detection algorithm lets designers use an industry-standard microcontroller rather than a more expensive digital-signal processing chip.
VCS directly measures the position of the pupil by using the retroreflection of infrared light from the rod and cone receptors lining the retina--much like the red-eye effect often seen in family snapshots where the flash is too close to the camera lens. An 880-nm LED furnishes low-power illumination, approximately 200 muW/cm2 at the eye. The light retroreflects from the retina back out of the pupil, off a mirror, through a beam splitter, and into a 384x287-pixel CMOS camera-on-a-chip from Vision, Edinburgh, Scotland. The entire system is small and light enough to fit in a comfortable headset and is compatible with eyeglasses. A wireless radio PC link is optional.
Low-cost, laser-generated holographic optics match the infrared incident beam from the LED to the camera field-of-view, producing a high contrast ratio between the retina-reflected light and that from the surrounding eye surface. This high ratio permits safe power levels without affecting performance. Optical filtering removes spurious visible-light signals. When the eye moves to a new point of interest, the camera tracks movement of the relatively sharp reflected-light beam and thus the pupil's new location.
Users calibrate the system by looking at and clicking on the four edges (top, bottom, and sides) of the area, or display screen, of interest. Depending on the application, users "click" by pushing a button or keyboard key. Alternately, staring or "dwelling" at a spot for a fraction of a second can cause the system to register a click. Calibration and driver software runs on Windows 95 and smooths out involuntary movements inherent in the eye. On a PC screen, a user can position a cursor within 1-cm-diameter areas at a typical working distance.
The first widespread use of VCS will be video gaming, predicts Gay, where mass production should bring its cost down to that of a high-end joystick. As a mouse alternative or supplement for hands-free control, the system is also well suited for aiding disabled people to communicate and function via PCs. He also cites telepresence applications such as bomb-disposal robots.
VCS's holographic optics and a CMOS video camera take advantage of the retina's property of retroreflecting infrared light to track eye position. |
Monitoring eye movement--as opposed to controlling something with the eye--finds applications in aviation training, to assess aircrew instrument scanning and observation of surroundings, and in market research to gauge product or image visual appeal. In medical schools, VCS can highlight a surgeon's focus of attention to medical students. Eventually, VCS could become a keyboard substitute for head-mounted computer displays.
All-new Malibu packs a punch
by John Lewis, Northeast Technical Editor
Pepperell, MA--With an estimated 4 million vehicles in the midsize-car market by the end of the century, engineers and designers must tune into the wants and needs of car buyers before turning to computers and clay models. While roominess, fuel economy, and modest pricing are still crucial to vehicle success in this segment, studies reveal increased consumer interest in driving enjoyment, new technology, functionality, and attractive styling.
Tough competition led engineers to benchmark more than 250 aspects of midsize cars before designing the Malibu. With these data in mind, Chevrolet set out to create a four-door sedan to attract both domestic and import owners. Result: a roomy, energetic sedan that offers many standard features usually associated with more expensive cars. Although the Malibu starts at $15,995, the LS model evaluated here has $18,319 on the bottom line. Not a bad price, considering these extras:
Custom reclining bucket seats, split folding rear seat, and luggage-area cargo net
Color-keyed front and rear mats
Electric rear-window defogger
3.1-liter SFI V6 engine
AM/FM stereo with CD player
Power door locks, windows, and mirrors
Inside rearview mirror with dual reading lamps
Remote keyless entry
Electronic speed control with resume-speed option.
While LS options are extensive, the base Malibu sedan includes air conditioning, tilt steering wheel, ABS, a four-speed automatic transmission, AM/FM stereo radio, dual airbags, and Passlock theft-deterrent system. In addition, its aluminized-stainless-steel exhaust system, two-sided galvanized exterior body panels, and up-level carpet and fabric may improve the car's appeal to practical, value-oriented people. That's probably one reason GM plans to make the new midsize entry its highest-volume sedan.
Speaking of volume, the Malibu is deceptively spacious. Front and rear, head and leg room is more than adequate. And my wife fit over twenty full paper grocery bags in the trunk. Other bonuses include:
Rear-seat heat ducts
Covered auxiliary power outlet
Backlighting on major switches
LEDs instead of bulbs in the instrument panel
Dash-mounted ignition switch
And a retractable cup holder located in the instrument panel just to the left of the driver.
Engineers improved the ride and handling as well. By specifying aluminum for the front and rear suspension knuckles, calipers, and wheel cylinders, they reduced the suspension's unsprung mass to give a nimble driving experience. A hydro-formed front frame combined with a stiff chassis gives a solid and stable ride. According to Tim Blenman, GM P90-platform structural engineer, the Malibu's static stiffness is 12.4 kNm/degree with three shock towers constrained. A solid foundation helps minimize vibration and provides a good platform for mounting other vehicle components.
Other NVH improvements on this sedan include heat-expandable seals within the framework, "roof deadeners" baked onto the roof panels, expandable foam between steel layers on the tunnel section of the floor, a double-wall steering boot that minimizes noise transmission through the steering column, and hydraulic engine mounts that absorb and dampen engine movements and vibration.
The only maintenance this vehicle requires for the first 100,000 miles is regular oil changes, tire rotation, and tank re-fueling. The bottom line is that Malibu offers surprising value for the money. Its ride, handling, and standard features should help Chevrolet fill a key product niche between the Cavalier and Lumina sedans.
New device offers fast switching, low power losses
by Charles J. Murray, Senior Regional Editor
Zurich, Switzerland--Hailing it as an "exponential technological leap," engineers from ABB Industrial Systems have unveiled a new power switching device that reportedly offers fast switching and low power loss. The firm's engineers describe it as simpler, faster, more efficient, and less costly to operate than conventional silicon-based power switching technologies.
Known as the Integrated Gate Commutated Thyristor (IGCT), the new semiconductor technology makes it easy and cost effective to switch up to 10 megawatts of power at medium voltages. Designed for high-speed switching, IGCT could serve in industrial motor drives, traction motors, ship propulsion, transmission lines, and many other applications.
"We liken this to the quest for the Holy Grail in power electronics," notes Eric Carroll, industry business segment manager for ABB Semiconductors AG, Switzerland. "It may not be the perfect switch, but it has taken us very close to that goal."
Silicon-based switches similar to the IGCT have long been used to control electric motor speed. Up to now, however, the power silicon switches have had difficulty switching megawatts of electricity at medium voltages (2.3 to 6.9 kV). The two most likely candidates for those applications have been Gate Turn-Off (GTO) thyristors and Insulated Gate Bipolar Transistors (IGBT). Still, both of those technologies have exhibited drawbacks in high-horsepower applications. GTOs are comparatively slow, and IGBTs, while faster, have larger switching losses. Many engineers believed that IGBTs were the future for such applications, but anticipated developments haven't happened yet.
With the introduction of the IGCT, however, ABB engineers say they have licked those problems. They employed a transparent anode design, which enables charge to quickly flow out of the device, enabling fast switching. They also reduced the thickness of the silicon wafer on which the device is fabricated. This permits the diode and switch to be fabricated on the same wafer, limiting the amount of stored charge. Result: Low-loss switching can occur at the device level.
With IGCTs, designers can now choose the switching rate to match their application requirements. The device's unique characteristics also enable designers to eliminate the need for so-called "snubbers" and large, complex inverter designs. Additionally, since IGCTs cut construction costs through thinner silicon wafer design and simplified power circuitry, the cost of power electronic systems can be reduced by 30%, ABB engineers report. Equally important, component cost can be reduced by 50%.
"A lot of people thought that IGBTs would eventually be better for medium voltages," notes Hal Stillman, vice president, technology and innovation, ABB Ltd., Zurich, Switzerland. "But our news is that IGCTs are better for medium voltages--and they're available now."
Welding helmet with a view
Wadenswit, Switzerland--The EagleTM Shade Master helmet totally covers the face, neck, ears, and throat. And, since it weighs only 18 ounces, filter included, it can be worn for long periods of time. Helping make this possible is the use of Ultramid(R) nylon, supplied by BASF Plastics (Mt. Olive, NJ). The material lets the company mold the helmet's walls only 0.047 inch thick.
A.C.E. International's Acromater Div. (Walpole, MA) markets the helmet in North America. Acro-Matic Plastics (Leominster, MA) molds them. "Molding the thin-wall helmet was a challenge," says George Doumani, VP engineering for Acro-Matic. "Good esthetics are important. The black helmet has a fine, pebble-like surface, and it's free of mold lines."
Doumani reports that his firm looked at many grades of nylon before selecting the impact-modified Ultramid resin. "We found that the resin reliably fills the thin-wall mold, and it's free of gassing, which could result in surface blemishes," he adds.
Tests conducted by Xelux Opto-Technics show that the helmet has good impact resistance, is unfazed by hot metal embers, and meets ANSI Standard Z87.1.
Resin scores with hockey mask manufacturer
Mississauga, Ontario, Canada--Hockey goalies are constantly under the gun as they battle to protect their territory from the invasion of a hockey puck traveling at 100+ mph. To maximize mask performance, Rhino Sports Group engineers chose Allied Signal's Capron(R) HPN 9315G resin, a nylon-based polymer impacted with rubbers.
"The Allied Signal resin is tough and lightweight, so players can wear the masks virtually without noticing them," says John Liddel, president of Rhino Sports and himself a hockey player. "The mask has an outer face protector and an inner padding insert, which all together weigh just two pounds."
Liddel also notes that the resin is easy to mold. Instead of using the traditional hand-made fiberglass process, Rhino injection-molds the resin into the mask. This eliminates secondary operations such as painting and sanding. In addition, molding reduces the weight of the plastic from 1,316 to 746g.
Rhino added ribs in critical parts of the hockey mask to increase stiffness. "Players like to have a mask that feels stiff even if it's lightweight," explains Liddel. Conventional fiberglass masks need a 55% glass reinforcement for consistency in molding. These masks are stiffer and have a rough surface.
A 60-second cycle time allows for the production of approximately 1,300 masks over 24 hours, up from the daily production of 25 masks by the hand-made fiberglass process.
Capron HPN 93156 maintains its inherent chemical resistance, particularly to grease, oils, and hydrocarbons. Other properties include a tensile yield strength of 22,000 psi, flexural strength of 32,000 psi, and a drop-weight impact of 40 inch/lbs. It is available in natural, heat-stabilized, and pigmented versions.
"We're using this advanced material to move the goalie mask market out of its conventional manufacturing process," Liddel says, "which dates back to the earliest days of hand-make fiberglass products."
Sensor housing subdues food chemicals
Gerlingen, Germany--Designing a special temperature sensor for the treatment and handling of biologically sensitive media proved a challenge for Endress+Houser Conducta GmbH & Co. engineers. For one thing, the sensor housing for the company's new CLS 51 conductivity measuring cell had to have a biocompatible surface and resist corrosive media. They found that material in a glass-fiber-reinforced polymer with a balance of chemical, mechanical, and thermal properties.
A conductivity measuring cell measures the electrolytic conductivity of media in the food and pharmaceutical industries. Fitted in line with standard adaptors, the device's housing must keep the sensor in place in the pipe while protecting it from attack by media flowing through the pipe. Applications include: separation of product/water or product/product mixtures; control of acid and alkali concentrations; and monitored control of cleaning processes and systems, such as CIP (clean in place).
Integrated into the crevice-free housing design of the CLS 51 cell is a temperature sensor characterized by ultrafast response to media temperature changes. Use of PEEKTM polyaryletherketone, supplied by Victrex USA Inc., West Chester, PA, resulted in a hydrodynamic exterior housing with inherently high surface quality. Because it is semi-crystalline, PEEK polymer has inherently good wear properties in most environments. It is also insoluble in common solvents, and resists a range of organic and inorganic liquids.
Adding glass fibers increases not only the chemical resistance of the base resin, but the material's mechanical strength at elevated temperatures. Tests show the polymer can withstand temperatures in excess of 480F for thousands of hours in steam or high-pressure water environments. Result: The CLS 51 passed with flying colors the rigid U.S. and European hygienic standards required for this device.
Terafloppers promise engineering revolution
by Walter S. Wingo, Washington Editor
Washington, DC--Engineering will never be the same now that ultracomputers--with their mind-boggling computational speeds--are here. So predict the people who run Option Red, the world's fastest computer.
Option Red takes up about 1,600 sq ft of a building at the Sandia National Laboratories in Albuquerque, NM. More than any other laboratory of the U.S. Department of Energy, Sandia is oriented toward engineering research.
Capable of performing 1.8 teraflops, or 1.8 trillion floating-point operations per second, Option Red last December became the first "teraflopper."
At a conference in Washington, DC, entitled "Revolution in Engineering," Norman Augustine, CEO of Lockheed Martin, which operates Sandia, described what terafloppers mean to engineers.
"Teraflop computers are changing the very concept of engineering," Augustine said. "We used to do tests; now we use models. It's a totally different system."
With high-performance computing, other Sandia officials explain, designers can create new products without the expensive, time-consuming physical prototype process. In its place: virtual testing of proposed products in a fraction of the time it takes using the traditional trial-and-error process.
Terafloppers also will provide algorithms for tackling broad, extremely complex problems. Engineers will be able to plug these algorithms into their workstations to help solve their particular design problems.
To handle such feats, Option Red is composed of 9,072 Pentium Pro processors with nearly 600 billion bytes of memory. The chips reside in 84 computer cabinets.
Four more terafloppers are planned under a 10-year government program to develop higher-resolution, 3-D physics modeling for evaluating an aging nuclear stockpile without actual testing.
The second teraflopper, Option Blue, is scheduled for demonstration at Livermore Laboratories in California in December 1998. It will be a 3-teraflop system. The remaining three machines, due in stages up to 2005, are to be capable of 10, 30, and 100 teraflops.
Time sharing. The key to tera-scale speeds is massive parallel processing. It breaks up a numerical problem so different clusters of shared-memory processors solve pieces simultaneously.
Though the main job of Option Red is to simulate nuclear tests, it will also provide a platform for commercial applications, including automotive and aerospace design, crash simulations, drug analyses, oil and gas exploration, and weather modeling.
Terafloppers can simulate a proposed product through its entire life cycle--from manufacture to disposal, decreasing defects while cutting costs in half, says Sandia official Russ Skocypec. It can speed up the manufacturing process from months to days.
Harold D. Morgan, manager of Sandia's Engineering Mechanics & Material Modeling Department, describes how high-performance computing solved a sticky heat-transfer problem with a brazing furnace for neutron generator devices at Sandia: "We did in a couple of minutes what would have taken more than four days on an ordinary workstation," he crowed.
While attacking such challenges, Sandia is developing techniques for making terafloppers more user friendly. Advances are reported in data visualization, networking, information surety, and mesh generation. Among developments is finite-element software for use on high-speed parallel-processing computers.
Partners sought. How can commercial engineers take advantage of the government's terafloppers? Sandia hopes to form new partnerships for collaborations and funding through a variety of arrangements. Among these are cooperative R&D agreements and licenses to permit Sandia to transfer intellectual property, including algorithms, to industry.
Confidence in the validity of teraflopper modeling and simulations is the key to its acceptance by engineers, Skocypec notes.
"The foundations of these simulations need to be verified," says an official of the National Science Foundation (NSF). "Otherwise you just get wrong results faster."
NSF and the Department of Energy will spend $2 million in the coming fiscal year to promote research into the engineering on which such complex computational analyses are based.
With computers already outperforming chess champions, will terafloppers eventually replace engineers? Quite the contrary, assures Sandia's Skocypec. Terafloppers, he says, are superb but mindless tools that permit designers to test the limits of their imagination without high cost or risk.
Wire alternative cuts soldering time
Marietta, GA--By replacing flux-core wire with a new type of paste solder, Servicetrends Inc. reduced the time and operator dependence of a precision soldering operation.
Servicetrends specializes in refurbishing the electrodes used in lithotriptors--machines that pulverize kidney stones with sound waves so patients can avoid surgery. In each unit, the electrode's conical upper tip serves as the hub of a "cage" of six insulated copper wires. After the tip has been removed and reground, the ends of the wires must be resoldered into six 1.35-mm-wide by 12.7-mm-deep holes in the brass outer conductor.
In the past, highly skilled operators used irons to force wire solder down inside the small blind holes. The process involved a significant amount of rework because excess solder had to be removed and any melted or discolored insulation trimmed away.
Electrode Department Manager Earnest Pringle believed he could cut process time and improve productivity by applying an identical amount of solder to each hole. He contacted EFD Inc. (East Providence, RI), which recommended SolderPlus paste solder.
Unlike wire solder, SolderPlus can be easily placed at the bottom of the small holes. A precisely timed air pulse from automated dispensing equipment--instead of subjective operator judgment--controls solder amount to ensure consistent deposits.
The new approach has enabled Servicetrends to reorganize the entire soldering process for greater efficiency. Previously, the wires had to be soldered one at a time. Now, the dispenser places an equal amount of SolderPlus in each of the holes in advance, and a microtorch reflows all six deposits at once. Throughput with the new method is much faster, and rework due to overapplication or scorched insulation has been eliminated.
"Our new method is quicker and more reliable," says Pringle, "and the more-consistent deposit results in a better-looking assembly."
What this means to you
Precision components that withstand higher loads, temperatures, and pressures.
High-speed machine tools requiring less coolant.
About the Author
You May Also Like