A specialist in the production of hydronic-based components for use in heating and cooling applications across the building spectrum--residential, commercial, industrial and institutional, Taco Inc. (Cranston, RI) manufactures pumps, heat exchangers, expansion tanks, flow measurement devices, air separators and valves, and zone control products.
Manufacturing engineers Bill Young and Joe Hutchins faced the challenge of higher production levels and broadening product lines at Taco. They were convinced that only by being innovative could the company stay ahead of its competition.
As a result of Taco's push for innovative products, sales had doubled in the last ten years. In the machining area, the manufacture of circulator pump casings was being performed on a rotary dial machine purchased in 1993. Its capacity and flexibility were being challenged. While it had the potential capacity to produce 700,000 parts per year, it was difficult to keep the spindles turning.
There were three sources of downtime for the dial machine. Changeover from one part number to the next took as long as an hour. If that were to occur once a week, it was acceptable. Young and Hutchins knew, however, that more new products were coming. They had seen the circulator pumps expand from twelve models in 1992 to thirty-five today, and the trend would not change. Most models of pumps were offered in cast iron or bronze. It was important to have quick changeover from part number to part number.
New products also required time on the dial machine to prove-out new tooling, workholding, part programs, and fluids. When Young or Hutchins engaged in prove-out, production went to zero.
Mechanical problems on the dial machine also increased downtime. The rotary's builder had gone out of business, and spare parts were difficult to obtain. In fact, when the company needed components to repair the rotary, Taco would acquire the part prints and manufacture them.
In addition to production downtime, there were a number of other issues motivating Taco to add new, flexible capacity. At times a finished part would reach final assembly and, despite a series of visual inspections, the volutes inside the casing would be clogged by casting-process debris. This problem had to be eliminated. The machined part also had to be cleaned and dried prior to delivery to the assembly area.
Young and Hutchins also had to address part cost. If the labor content was too high, Taco would have to look at offshoring the parts. This was not the company's desire, but a competitive market required them to control costs. Sales growth had pushed Taco into an overtime situation, and their two shifts were sometimes working 10 or 12-hr days six days a week. Taco management wanted to eliminate the overtime and minimize labor costs.
One further constraint was floor space. Taco's president, John White Jr., was determined not to add bricks and mortar. He insisted that the manufacturing team raise the productivity of the machining area, yet fit it inside the existing four walls.
| HVAC Success Strategies
Taco Manufacturing Engineer Bill Young carries in his pocket a card containing the nine strategies that guide his company. Developed in 1998 by Taco President John White Jr., the grandson of founder Elwood White, these strategies outline Taco's vision for the 21st century.
One of them, innovation, has driven the company to broaden its product lines, and inspired manufacturing personnel to develop new flexible manufacturing systems to produce these new products. As innovation spawns new products, it also changes the game on the shop floor. Flexible automation becomes the optimal solution to high-volume manufacturing.
| Taco's Strategic Plan
Financial Stability and Growth
Human Capital Development
Founded in 1920, Taco has a workforce of about 500. Its products are sold throughout North America, South America, and Asia, and annual sales exceed $130 million. Three generations of Whites--grandfather, father, and son--have been active in the company.
Elwood White, a former radiator salesman, started the company in Elizabeth, NJ, when he purchased the Thermal Appliance Company. The company soon relocated to Providence, RI, after Elwood White teamed up with inventor Robert Blanding. Its first successful product was the tankless water heater, which was equipped with a second innovation, a tempering valve to mix hot with cold water.
Facing these varied challenges, Young and Hutchins put together a four page request-for-proposal (RFP) that was sent to several machine tool builders, distributors, and systems integrators. Their vision of how to be successful was what they called a Partnership Turnkey. Key suppliers would collaborate with Taco and one another to deliver a complete solution to Taco. After numerous responses, a solution began to emerge.
The part volume requirement of 500,000 parts per year---combined with limited floor space--pointed the manufacturing team toward some form of automated cell. Young and Hutchins met with a number of system-integrator and automation companies, and one--Wes-Tech Automation (Buffalo Grove, IL)--emerged as a preferred supplier. Wes-Tech worked with Taco on a series of cell concepts, including both three-axis gantry loaders and six-axis articulated robotic arms to load and unload HMCs. There was much discussion about the gantry loader versus the articulated arm. Personnel at Wes-Tech and Taco knew there was a need for precise casting placement and auxiliary operations such as air blow-off to rough-clean and dry the part. The articulated arm's flexibility lent itself to performing these functions.
Rail guided vehicle (RGV) pallet-shuttle systems were reviewed, but they required more floor space than the team could spare, and high fixturing costs ruled out those proposals. Despite a higher purchase price, the articulated arm was chosen. Wes-Tech assisted in these discussions, helping the Taco personnel weigh pros and cons.
Once the overall cell configuration was determined, the requirement for flexibility turned the discussion toward a system to properly present raw parts to the articulated arm. Once again, Wes-Tech was able to help Taco by offering a dual-track conveyor system that enabled the cell to manufacture two different part numbers at the same time. The conveyors contain up to twenty pallets, each holding one part. Wes-Tech was able to design the pallets to cradle as many as thirteen different circulator cases, and future casing designs would fall into this constraint. This approach kept investment and maintenance of the conveyor system to a minimum.
The cell design called for an operator to load the raw casting onto the pallet. Then the pallet passes through a Poka Yoke (mistake-proofing) station designed into the conveyor system. Poka Yoke is a lean manufacturing technique that prevents the improper assembly or processing of a component. This simple mechanical device uses a mounted switch to determine if the part has been improperly loaded into the pallet. If the part is loaded incorrectly, the switch stops the conveyor immediately, preventing it from presenting an improperly fixtured part to the arm, avoiding a potentially expensive wreck.
Once the raw casting clears the Poka Yoke station, it enters a photo-based imaging system that checks for clogged volutes. By shining a bright light into one end of the casting and measuring the amount of light and its pattern transmitted through the other end, the inspection station can determine if there is a clear passage through the casting. If the passage is clogged, the arm places the casting in a reject box.
After the part is inspected and enters the cell, the overhead articulated robotic arm becomes the center of action. In the process of selecting a supplier, Taco and Wes-Tech knew that the most critical factor for the arm was reliability. While comparing different suppliers of articulated arms, Young and Hutchins visited a die-cast foundry that was using GE Fanuc arms (Fanuc Robotics North America, Rochester Hills, MI) to load and unload castings. Bill and Joe were convinced that if the arm could operate in a foundry environment it could certainly perform in their machine shop.
The four HMCs in the cell would be tooled to machine any of thirteen part numbers, allowing Taco to manage production mix simply by selecting the part program in the HMC. Managers could use all four HMCs on one part number or adjust the mix to run two parts as needed. Once they selected the proper part programs, the arm's actions were slaved to the actions of the HMCs.
When a part is finished, that's the signal to the arm to grab a green part off the proper conveyor, move to the HMC, and exchange green part for finished part. The arm then places the finished part on the output conveyor and returns to its "wait" position until the next HMC calls to be loaded.
After the machined part is placed on the output pallet, it moves along the output conveyor to a part-washing station. Taco had experience with a wash station that worked very effectively on its dial machine. When they explained this to Wes-Tech, Wes-Tech was able to recommend a company that had designed and delivered a number of successful washing systems to Wes-Tech. This part-washing system company, Alliance Manufacturing Inc. (Fond du Lac, WI), was willing to travel to Taco's factory, study the existing washer, and design and build a new system for the circulator-casing cell.
There were a number of potential suppliers of horizontal machining centers. The Tyler Machine Tool Co. (Seabrook, NH), a local machine tool distributor that had a long working relationship with Taco, offered the Mori Seki SH Jr. (Mori Seiki USA, Irving, TX), a very compact, ram-style HMC that met Taco's floor space constraints but was not typically offered in the US.
Bill Young was determined that this flexible circulator casting manufacturing cell must go in and be up and running immediately. The company could not afford a long installation, debug, and ramp-up period. Suppliers all agreed to set up the complete cell in Tyler's tech center in Seabrook. Taco would ship castings to Tyler, and the system would be proved out before hitting Taco's floor.
The runoff allowed engineers to identify and resolve a number of problems. For example, chip conveyors were not built to spec, and would have dumped chips onto the floor rather than into Taco's in-ground flume system. As a result of this prove-out at Tyler, the cell was up and running full production at Taco two weeks after delivery.
Taco also improved part processing using two key techniques. Working with PowerHold Inc. (Middlefield, CT), Young and Hutchins were able to design a fixture that enabled them to process parts complete in one setup. The fixture contained a through-hole in the center that allowed drilling of four mounting holes from the back of the part.
The two manufacturing engineers also learned that a new drill from Mitsubishi Materials USA Corp. (Rolling Meadows, IL) allowed them to increase drilling feed rates enough to drill four holes with a single spindle faster than they had been able to drill using a multispindle attachment.
The new system meets production goals, and allows Taco to maintain 75% of production output when one machine needs to go offline for new part prove-out or maintenance. Young and Hutchins say there were several keys that made the project successful:
- Communication of requirements--the four-page RFP that the manufacturing team created was a document that clearly communicated Taco's expectations, as well as concerns.
- Selection of proven partners--an experienced automation company and machine distributor provided the help Taco needed to purchase and install its first automated cell.
- Be innovative and open to new ideas--in a number of instances the Taco manufacturing team went into a discussion with preconceived ideas on how to proceed, but they listened to their suppliers/partners and ended up with a superior process.
This article was first published in the June 2004 edition of Manufacturing Engineering magazine.