"No Gear Is an Island," Feature Article, June 2004

no gear is an island

A system is the sum of all its parts, both good and bad.

by Steven M.
Zychinski

Most people probably don't give a second thought to the works of a vending machine, as long as it accepts their money, returns the correct change, and gives them what they paid for.

When someone deposits money into a vending machine, the bill acceptor or coin changer validates it as legal tender and issues credit. Once a product selection button has been depressed, the control system tells the machine what product to dispense and tells the coin changer to issue the correct amount of change.

Coin Acceptors Inc. designs and manufactures vending machines, control systems, bill acceptors, and coin changers for the snack and beverage industry. The company, based in St. Louis, sells its products through branch companies as far away as Europe and Australia.

To improve the industry's machines, the firm set out to offer a high-capacity coin changer, which would operate a bit differently. It was during the design of this device that our engineering team at Coin Acceptors was reminded of an important design principle: No gear operates alone.

Our explorations to come up with the right design even included a modified cut of gear teeth to avoid problems of excessive contact.

Small gear trains (above) rely on carrier plates about the size of a quarter to maintain alignment. The plates are secured with center-punched pins (below).

A coin changer usually has a supply of coins stacked in tubes. The new design would collect the coins in a hopper, and let a mechanism at the bottom orient them for payback. It would give the coin changer greater change-making capacity.

A larger change capacity would improve the productivity of the company's machines. The more money a coin changer can store and dispense, the more profit the vending machine can make for its owner, because it can sell more merchandise between routine service visits.

There were some special considerations in the design. Since a significant percentage of vending machines are used outdoors, they are subject to a range of environments. The company sells its machines in countries around the globe. One machine can be selling sodas on a street in Quito, Ecuador, almost on the Equator, while another may dispense quick-energy snacks outside a filling station in Kiruna, Sweden, above the Arctic Circle.

The company determined that, to function reasonably in the extremes, the system needed to operate at ambient temperatures between -10°F and 160°F. Due to the nature of the coin changer's dispensing mechanism, it had to output a large amount of torque with limited available power, and had to be rugged enough to reverse out of hard stops 10 percent of the time in the course of its lifetime.

The stops were a consideration because of the hopper system of storing coins. The sorting mechanism included an auger system that oriented the coins. Sometimes there would be a blockage of coins in the hopper. The auger was designed to persist for a fraction of a second when it met a blockage. Then it would reverse itself and restart.

When the coins fell in line, the mechanism would dispense a coin every half second until the total sum of change was delivered. If there was a blockage, the coins might take a full second each to drop.

The engineering team at Coin Acceptors decided to outsource the drive system design to a contractor.

Because the coin changer had to fit existing machines, its exterior dimensions were already settled at approximately 14.5 inches in height, 5.5 inches in width, and 3.0 inches in depth. The motor selected by the contractor left less than a half-inch of height available for the gear train, so a fixed differential epicyclic drive train was selected for its robustness and reduction capability within the allotted package size.

Lubricant was used to reduce heat buildup and increase the efficiency of the system. The drive train's planet gear carrier was an assembled steel structure to maintain gear alignment. All gears were injection molded out of unfilled acetal plastic, except the sun gear, which needed the increased strength of C36000 brass. Injection-molded plastic gearing offered more design freedom, quieter system operation, and lower cost.

What We Like to Call 'Challenges'

We subjected the first prototypes to stringent testing from -10°F and 160°F. To say they displayed poor performance is an understatement. Low output torque coupled with drive system lockups plagued early test efforts. It turned out that every component contributed to the shortcomings of the system. The motor was noisy and weak, the lubricant hampered the system at low temperatures, the carrier deformed, and the gears were misshapen and suffered backlash that increased with the heat.

With the production date for our coin changer looming, we needed to come up with solutions fast. The company decided to bring the design in-house.

Since there was a hefty investment already made in tooling, inspection equipment, and assembly fixturing, using any of the existing hardware would be beneficial.

We started with the motor, since a gear train is only as good as the component that drives it. The 12-volt dc motor squeaked and drew more than twice the current we had expected under no-load conditions at -10°F. The internal bushing of the motor was causing additional loads as the ambient temperature dropped.

We found another model from Mabuchi Motor Co., which has a U.S. office in Troy, Mich. The motor has better performance characteristics throughout our temperature range, and within the same package size. It can output 40 grams-centimeters of torque at 9,000 rpm with 4 watts of power.

The coin changer will serve to keep exchanges in order in machines like this demonstration model. The gears must work everywhere, from the Equator to the Arctic.

Because the original lubricant was a high-viscosity hydraulic and circulating oil, we expected a slight reduction in drive system torque at -10°F, but actually lost more than 50 percent. Since the drive system operates intermittently, as it sorts coins only when change needs to be made, starting torque rating for the lubricant is crucial. The lubricant was actually decreasing the efficiency of the system.

We replaced it with a light-viscosity, synthetic hydrocarbon grease made by Nye Lubricants, which is based in Fairhaven, Mass. Efficiency loss of the system at -10°F due to the lubricant dropped to 5 percent. Since the drive system is not sealed, the grease seemed to stay with the gear teeth longer than the oil had done, giving the added benefit of extended life.

The planet carrier originally was assembled by press fitting steel pins into steel plates to form a rugged structure for maintaining gear alignment during a hard stop. The steel pins gave the planet gears a smooth bearing surface to rotate about and the steel plates were supposed to keep everything aligned.

When we tested them, the carrier plates separated from the pins and compromised gear alignment. Stress calculations along with focused testing confirmed that the plates were not thick enough and would not be able to handle the required loads.

Since increasing the plate thicknesses was not practical, a secondary operation was added to the assembly fixturing for a minimal tooling charge. Center punching the pins achieved the maximum holding force through a riveting effect and increased the rigidity by 50 percent. This allowed us to use existing plate and pin components.

Then there was the gear train and its array of errors in both design and processing. The design offered poor load transference with low contact ratio and excessive backlash, which grew worse as temperatures rose. It was clear the gear train, as designed, would not be usable in our specified temperature range. Combine that with the lack of adequate bearing support to maintain center distances and the runout associated with low-quality gearing, and you get a recipe for tooth wear and fracture, and ultimately drive system lockup.

Load analysis suggested a benefit from a material change for the ring gears. We chose a reinforced nylon from Ticona of Summit, N.J. The material offers more strength without increasing wear. It also has a reduced coefficient of thermal expansion, which meant 25 percent less fluctuation in backlash and contact ratio throughout the wide operating temperature range.

Undercutting planet teeth avoided problems, including high wear, that would come from too much contact below involute.

Inspection information confirmed that we had a quality issue with the ring gears. Their oblong shape, due to irregular shrinkage, was evident in the amount of runout measured. Higher-precision tooling was built and we took this opportunity to design the components to be more symmetrical, to obtain uniform shrink after being ejected from the mold. The new material was also beneficial here, since its mold shrinkage characteristics are 75 percent less than those of the original material.

You can have gears of the highest quality, but if alignment isn't maintained, there will be problems. We optimized all bearing supports and all but eliminated any loading applied directly to the motor shaft. This shifted the contact surfaces to the components with the highest output of torque.

As for the gear profiles, there were two schools of thought about allowing operation throughout the temperature range while using the existing carrier center distances.

The more conventional method meant increasing the diametral pitches of each gear. Adjustments in tooth thicknesses and heights would increase the contact ratios and reduce backlash, but we feared that the finer teeth of the ring gears wouldn't be able to withstand the hard stop requirement of the drive system, especially at 160°F. In addition, new inspection equipment would be needed from the diametral pitch change, and we were not very excited about spending that time and money.

We Can Do That?

We chose a more progressive method, and stuck with the original diametral pitches of each gear.

Adjustments in tooth thicknesses and heights increased the contact ratios and reduced the backlash between the gear sets, but allowed teeth on the ring gears that were thick enough to withstand the hard stop requirement of the drive system. It was necessary to alter the profile to avoid contact below the involute, which would cause high wear, high heat generation, low efficiency, non-conjugate motion, and noise. So we undercut the planet teeth. Now, no matter how much the gears expanded or contracted throughout the temperature range, they would still be maintaining adequate contact ratios.

Output torque of the drive system at temperatures between -10°F and 160°F proved to be at the specification. Life testing of the drive system at the full range of temperatures proved the hard stop performance to be 33 percent above the specification, or four lifetimes of the coin changer.

With these results, our coin changer was able to meet its production date. Coin Acceptors introduced the finished product to the market in August last year.

The challenges that we faced with our drive system are not uncommon ones. However, the way you face the challenges makes the difference. We always have to keep in mind that a system is made up of more than just gears.

Steven M. Zychinski is a senior engineer at Coin Acceptors Inc. in St. Louis.