Picture this: Your four-year-old grabs the camera off your dresser and decides to play photographer.

"Click."

The floor, the closet, the bottom of a chair, the TV, and the dog's tail, all become subjects of a diminutive domestic documentary. But except for the subject matter, every picture comes out perfect.

A budding Ansel Adams? Maybe, but credit these photos to Kodak's version of the Advanced Photo System, known as Advantix.

Unveiled in March, 1996, and officially available in May, the Advanced Photo system (APS) is the most massively engineered and funded line of photographic products in history. Not simply a slightly different camera or tool for glossier prints, it includes a new film base, emulsion, transparent magnetic coating, cartridge, cameras, loading interface, negative size, print formats and photofinishing equipment--all designed to work together. With APS, engineers have turned what was too often a frustrating experience into literal child's play.

Among their key mechanical innovations:

Creation of a 100% recyclable film-cassette thrust "engine", designed to six-sigma quality standards.

Brand new film. Engineers also created an entirely new film made from A-PEN, annealed polyethylene naphthalate, a polyester commonly found in magnetic tapes. A-PEN is flatter, stronger, and at 0.110 mm, 30% thinner than the cellulose acetate that's been the material of choice since it replaced flammable celluloid in 1924.

The film is narrower, too. Negatives measure 24mm wide instead of 35mm. Their 40% smaller size allows for more compact cameras. And, since the film sits closer to the lens, more light hits each square millimeter of film. This permits slower, finer grain emulsions to be used that produce equally sharp prints from the smaller negative.

A transparent magnetic coating, 1/100th the density of that on video tape, covers the back of the film. It allows, theoretically, about 2000 bytes of data to be stored on each frame. For now, just 400 bytes are used to catalog such things as the date, time, ISO, shutter speed, film speed, f-stop setting or if the shot was taken in artificial light.

Downstream, in the processing cycle, photofinishing machines use the data to correct lighting and exposure problems. And, after processing, consumers will be able to use any of hundreds of new APS devices hitting the market that leverage the technology. These products will treat pictures like data to be scanned, cataloged, printed, viewed on TV, manipulated on a PC, or sent over the Internet.

International team. Seventy-six critical patents form the heart of the APS package. Kodak owns 43 of them.

Nevertheless, the project was truly an international effort. Five major players--Kodak, Canon, Nikon, Minolta, and arch rival Fuji--huddled together in late 1991 to lay out the ground rules for the project.

Over the next two and a half years, the team synthesized inputs from every portion of the photo industry into a set of standards and specifications by which they could all play. "It was essential, but tough, to get everyone to agree on a common feature set," says Terrence McArdle, public relations manager for consumer imaging. "This was an opportunity to avoid the kind of scenarios that killed DAT and Betamax because nobody could agree on a common format."

Kodak provided marketing analysis from its "Voice of Customer" project. In 45 studies, more than 22,000 consumers from 11 countries explained what they wanted from photography. High on their list were eight things:

Greater portability

Kodak and its partners delivered on all these requests. Here's how.

Styrene 'thrust engine' drives film. At the heart of Advantix lies the cassette. Unlike the 35mm cartridge it resembles, the cassette isn't simply a box that keeps out light and dirt. It's a little machine containing 11 parts that thrust film, communicate status information, store negatives, and interface with any of the hundreds of downstream film-reading devices.

One of the biggest headaches was the design of the film-drive method. To facilitate drop-in-loading, the cassettes are sealed, with no exposed leader. Thus the film must be driven not pulled from one side of the camera to the other. "How do you 'push the rope' those 180 mm against considerable friction?" Bill Atkinson, system manager for APS, asks rhetorically. "It was commonly felt to be not possible."

Engineers produced almost 20 prototypes of different thrust-engine schemes, trying to find the best combination of force, reliability, cost, and compatibility with other portions of the design. The winning design: two thin styrene drive disks that pinch the coiled film on each side. As the spool turns, a small plastic pick built into the case scoops up the leading edge of the film and guides it out the cassette door. The drive disks apply enough side pressure to the film to actually bow it slightly in the middle, an intentional effect that increases the film's stiffness.

The film can produce a surprising one to two pounds thrust. On rewind, one drive disk freewheels to prevent driving the film to the outer periphery of the cassette and causing a jam.

Even something as simple as opening and closing the film door gave designers fits. Three different factors contributed to the torque necessary to open the door, and camera manufacturers wanted the value as low as possible. At the same time, to prevent inadvertently exposing the film, the door had to stay closed even when the cassette was dropped from two meters onto concrete. Ultimately, engineers trimmed 0.25mm from one part, and reduced the rise of a tiny detent by a few thousandths of an inch.

CAD ties design and manufacturing. Not one component in the Advantix system has a tolerance greater than 0.005-in. "Several thousandths runout in the spool was enough to cause problems," says Kodak engineer Henry Freidhoff. He points out several 0.003-in details. And, in one case, he reduced a wall thickness by just 0.0003-in for clearance.

Yet more than any other factor, the cassettes had to be supremely, almost impossibly reliable. People expect the film to be good 100% of the time. And unlike 35mm cartridges, each might be used 50 to 100 times over its life. "Think of the film jamming in the camera during your trip to the Swiss Alps, " says Tim Durkin, film manufacturing manager. "That's a super catastrophic failure that could never be accepted."

To meet the requirement of six-sigma quality, engineers sent roughly 60,000 prototype plastic components through performance testing. These led to more than a dozen subtle mold changes to each of four complete families of parts. Even once the cassette design was finalized, testing didn't stop. By March 1, 1996, the camera design group had run more than 250,000 cassettes through 8,000 prototype cameras, with only four failures attributed to the cassette.

Parametric Technologies' (Waltham, MA) CAD/CAM package sped these design iterations. "What made this possible was Pro Engineer," says Freidhoff. "We could quickly go from the program's 3-D design package to a mold-flow analysis package." He was able to address draft and molding during the design and not have to hand off the CAD models to manufacturing. "We were turning around hard tools in less than two weeks," he says. Prior to production, technicians measured more than 1100 dimensions of the molds to confirm the accuracy of the parts.

For the camera design itself, and the photofinishing system, engineers used EDS's Unigraphics software. Before turning to CAD, however, engineers performed massive statistical analysis intended to ensure six-sigma quality. They applied the principles behind buzz phrases such a FMECA (failure mode effect in critical analysis), QFD (quality function deployment), and Taguchi (quality control methods). And they generated reams of charts, graphs, and spreadsheets analyzing most every possible variable. "We did these studies virtually everywhere in the design," says Durkin, paging through a massive report.

As an added burden, Kodak decided it also wanted the cassette to be fully recyclable--label, bar-code disk, springs and all. "So here we are, designing these different parts, and we had to consider whether we could produce each of them from styrene," says Durkin. They could. Today, old cassettes can be ground into material for new cassettes by simply stripping out the film.

Low-friction pad mates head to film. When a consumer drops a cassette in a camera, s/he launches the film on a tiny journey, dipping under the platen, passing in front of the lens, and squeezing through one unusual landmark, the magnetic head. Until APS, cameras didn't have magnetic heads because film didn't have a magnetic coating. Now it does, and the challenge of incorporating a head into the camera was formidable.

Film is about 4.5-mils thick--two orders of magnitude greater than recording tape--with a magnetic coating just 1/100th the density. It's also curly, giving engineers fits with getting it to conform intimately to the head.

Certainly, with sufficient force the head could get most anything to conform. But such pressure would generate wrinkles in the image area or offer too much resistance for the film to push through.

Paul Taillie, senior design engineer for the head/film interface (HFI), examined, modeled, and tested the problem quite literally for years, narrowing the contending designs to five. His favorite, the double-humped camel pad, tolerates large variations in alignment (0.017-in), easing manufacturing. The runner-up was a bubble-spring concept that achieves the best read/write performance--so long as the pressure spring and the head gap are aligned to 0.004-in. This high-precision design found its way into the expensive photofinishing equipment.

Taillie got the camel pad idea from Canon. They were attempting to make the pad of Kraton rubber coated with Teflon. But the leading edge of the film tended to skive the surface, limiting head life. "HFI design can kill a project," Taillie says. "Many times you don't even realize you have a problem until it's too late." Ultimately he found a low-friction polyolefin called Lubmer (Mitsui Petrochemicals) that withstands the camera's 100-roll life span. Some popular materials, such as Delrin, can't be used because they are photoactive.

A coil torsion spring acts on a lever to press the pad against the head with about 70-gms of force. A good portion of the load is used simply to flatten the film. The pad rides atop a pin that is exactly constrained by two opposing V-blocks formed in the camera case, one at the base of the pin and one at the top. "As you drag the film through the camera, the V-blocks automatically align the pad to the head," Taillie explains.

The head writes on one of the film's two narrow tracks. Fear of scratching the film with dust has kept the head out of the image area--for now. In a Kodak-famous "dust test," Taillie recruited employees to bring in more than 30 kinds of dirt, such as shaver stubble and baby powder, to see their effect on the film and head. Full-width heads are still a few years off. "We've got machines in the lab writing to the entire film surface," says Atkinson. "It will happen."

Closed-loop motion control system drives zoom-lens photofinisher. Over the staccato roar of a photo processing machine pumping out 16,000prints an hour, a smiling Chris Kralles details the engineering virtues that makes Kodak's $480,000 CLAS III Photofinishing System the fastest and most advanced in the world. "This machine is beating the shorts off our competitors," he says. And he should know. Under the hood lies an innovative lens assembly for which Kralles heads the mechanical design effort. It zooms, adjusts aperture, and refocuses every 250 msec, allowing a single CLAS III to rapidly switch among APS' panoramic, HDTV, and classic print sizes. "With any other system, you would have to pass the film through the printer three times, one for each format," he says.

His design consists of two lens barrels arranged on a common axis, and a separate diaphragm, or iris, that regulates the quantity of light used to print. The barrels ride on six 6-race linear bearings (Nook Industries, Cleveland, OH) gliding over two, parallel, Armoloy-plated rods (Armstrong Bros. Tool Co., Chicago, IL). To avoid over-constraining the design, Kralles used spring-loaded bearing mounts on one rod that only stop the lens barrels from rotating. Extensive use of photolithography allowed engineers to catch several interference issues with the mechanism before they became problems. In fact, Kralles used stereolithographic parts to create the production casting molds for the lens barrels.

The same concern for over-constraint applied to the lens drive system, where engineers invented and patented a unique mounting method. There, two Compumotor (Rohnert Park, CA) step motors attach to a flexure that allows axial and angular movements, but resists rotation. The motors drive the lenses via two Kerk Motion Products' (Hollis, NH) leadscrews with anti-backlash nuts. A pair of bearings near the base of each shaft support the leadscrews. Both bearings are preloaded axially in opposite directions to eliminate free play of the shaft. But, uniquely, one of the bearings is staked into its mount whereas the other is clearance fit. "This gives a drive mechanism that is exactly constrained," says Kralles.

Completing the drive is a closed-loop servo system that uses a glass scale encoder (Dynamics Research Corp., Wilmington, MA) to provide positional accuracy of 15-microns throughout the lens' 3.2to 10.6magnification range.

A second patent applies to the iris. It consists of six blades mounted with ABEX 7 bearings and powered by a moving-coil actuator. The actuator drives a 0.001-in thick phosphor- bronze band coupled by flexures to the blades. This provides exceptional accuracy and repeatability. "Over its life, its precision changes only one-half percent," says Kralles.

Cutting manufacturing effort and cost weighed heavily into Kralles design as well. By careful control and selection of dimensions and assembly methods, he created a lens system that falls within spec simply by putting it together. "Typically, optical systems need lots of adjustments," he says. "With this, we just drop the elements in, stake them with epoxy, and they are within tolerance every time."

Each CLAS III lens system is designed for 40-million zoom and focus cycles, and it's been tested to 80-million before failing. The iris, engineered for 200-million actuations, is still running in a test fixture after more than 1-billion cycles. "This mechanism maintains repeatability over the entire life cycle--nothing changes," says Kralles. "The quality of print one is the same as print ten-million-and-one." It's such attention to quality that has made CLAS III--and all of Kodak's APS products--truly world class.