Balance Shaft Machining Process Gets Lean

For Anderson Machining, that meant revamping balance shaft machining processes that had been in use for four years. The company's main goal was to produce complete parts in a single setup on a just-in-time basis. VP of Manufacturing Erik Anderson sought his own partner to tackle the makeover, and after some research and a recommendation from another shop settled on Fuji Machine America (Vernon Hills, IL).

---

Anderson Machining's setup for producing Harley balance shafts (clockwise from top): Overall view of five-machine cell

---

Anderson also felt Fuji's process-driven approach to automation and 25 years of experience using robot-integrated machine tools supported the company's take on lean manufacturing, which the company calls "Competitive Pricing Through Productivity." The method dictates that all operations are coordinated to produce a completed part, and that product is ready to be shipped as close as possible to when the costs of manufacturing are incurred, he explains.

"We make just what we need to in order to replenish what's shipped on a given day. With batch processing, if there's a problem, everything in the pipeline between the first operation and everything that's running in all subsequent operations becomes a problem. But, when you start with a raw part and it comes out of the line a finished part, all the quality functions are done at one time. It's either a good quality product at the end of the line, or it's deviant and then there's one bad piece as opposed to an entire bad batch. We can monitor process capability all the time as well," he concludes.

Anderson Machining's previous batch process for the two-pound forged steel balance shafts consisted of two rough turning centers feeding a central finish grinding process, followed by manual deburring. The operation ran three shifts a day with three operators per shift.

The revamped process includes five Fuji machines--two horizontal machining centers and three two-axis turning machines--with automation and material handling to link all components of the cell. It begins with fixturing of a forging onto a workstocker. Each machine in the new cell contains its own robot; each set of machines share a transfer station. A Fuji HM-30 HMC center mills the ends and machines in centers for subsequent turning operations.

The part is then transferred to a Fuji ANS-31T two-axis lathe for rough turning. Custom workholding locates off the center and clamps on the forging OD. After transfer to a second ANS-31T machine, the two end-bearing journals are rough turned and the required thread is rolled.

After automatic reorientation, the part moves to a third ANS-31P, a precision two-axis lathe for hard turning and finishing that replaced two grinders from the previous process. Using a face driver for workholding, this machine turns and burnishes both journals for finish.

This machine also features automated gaging of both part and tooling to determine if process adjustments are necessary. If necessary, the line shuts down and the part is removed and isolated. The machine then automatically compensates for tool wear based on the gage data and continues operation.

After another transfer, the workpiece is completed when a second HM-30 HMC mills a flat, drills a hole, and deburrs 95% of the workpiece.

According to Anderson, the process can hold tolerances of 11 µm on diameter, 12 µm TIR runout, and 12 µm on surface finish. The line also had process capability (Cpk) better than 1.67 in runoffs in Japan and on Anderson Machining's shop floor.

The new setup produces 56 balance shafts per hour, a 33% increase from the 42 pieces/hr produced by the old process, allowing Anderson to eliminate a shift while gaining capacity. Scrap has decreased from 1.5% to <0.5%. Seven of the nine operators required for the prior machining method have been relocated to Anderson's Whitewater, WI plant, reducing the number of operators for this process to only one each for two shifts. Annual production costs for the component have been cut by $400,000.

According to Erik Anderson, changing from stand-alone machines to cellular production also requires changing thought processes. "In the 1990s when we were a job shop, we ran like gangbusters on our individual machines, so we know a lot about machining fast," he says. "Back then it was only about cycle time. But when you start pushing machines together, you have to start thinking about takt time and line balance.

"For example, one of our operations takes 55 sec--longer than the rest. Other operations could take only 30 sec, but why bother beating the machinery? That operation's going to take 55 sec, so we slowed certain things down to improve tool life. In fact, our tool life has increased by ten times.

"Americans tend to run the heck out of machines," he continues. "When you're balancing a line like this, it doesn't make sense to do that. Takt time--the time it takes from when one finished part is produced until the next finished part is done--is what we care about the most. The volume of this product line dictates a one-minute takt time."

Erik Anderson says flexibility was another reason Anderson Machining chose to partner with Fuji. "Harley-Davidson is aggressive on updates," he explains. "We rarely run a part for two years without significant changes, and it was cost-prohibitive to update with other machine tool builders concepts." Anderson considered everything from conventional turning to automated turning cells, automated grinding, and dedicated transfer machinery before selecting Fuji's self-contained automation.

"When you're looking at stuff like this for Harley-Davidson, they want to know what you're going to do if the volume doubles; they want to know what you're going to do if the volume drops," Anderson says. "With this, I can say if the volume goes up, I'll add machines, and your price will go down. If this project goes south, the machines are versatile enough that we could split them up for use on other applications," he concludes.