Cutting Manufacturing Costs

Randy Heyler

Faced with rising costs, photonic component manufacturers must consider all available options, their benefits and trade-offs.

With increasing market pressure to make significant reductions in the cost of manufacturing photonic components, engineers are turning more and more to new design approaches, higher levels of device integration and the extensive use of automation for their next-generation device designs. But no single solution can relieve the ongoing price pressures. A look at some approaches to reducing cost could prove beneficial.

First, it is important to understand the base costs and cost drivers in any reduction effort (Figure 1). This example of a cost breakdown for a photonic component is divided into material, labor and equipment-related content. Note that both the material and labor portions include constituent costs related to scrap and rework — the principal penalties one pays for poor yield. To simplify this analysis, “overhead” other than equipment cost may be considered within the labor and material portions. Another way to express this is a simple formula: assembly cost = material + labor + scrap material + rework labor + equipment depreciation.

Figure 1. In this example, material and labor are direct variable costs. Scrap cost is material lost because of poor yield, and rework labor is cost to repair/refurbish or to salvage a bad component. Equipment cost is the fixed price of equipment, divided by the number of components made overits useful life.

Although this profile can be considered typical for several applications, preciseness is not critical. Rather, it is more important to understand the relative magnitude of the constituents to determine the greatest opportunity for cost cutting.

Cost-reduction efforts

It is also important to correlate each of the following five cost-reduction efforts with the category that will be affected, so that the potential benefits and chances for success can be determined.

Figure 2. Semiautomated assembly workstations improve efficiency and process consistency, delivering increased yield.

• Automation: The introduction of automation into any process typically improves efficiency and yield. Efficiency is a measure of the “assembly cost” in terms of nonmaterial expenses. These include labor, floor space, equipment and utilities on a per-part basis. Yield, however, is a measure of the “scrap and rework cost” of having an inconsistent or uncontrolled process, and it represents the cost of damaged material and additional labor and/or material needed to remanufacture or repair the part. Automation can achieve substantial improvement on both counts.

A look at the improved efficiency gained through introducing a semiautomated workstation on a typical butterfly fiber pigtailing process makes the benefits of automation obvious (Table 1). In this comparison of a manual vs. a semiautomated assembly process, note that, although capital equipment cost is higher, the per-part equipment cost remains unchanged, and throughput triples. This payback calculation is conservative, because it does not include potential scrap savings realized through improved yield. The cycle time and consistency improvements result in a lower per-part cost and a fast payback for the initial equipment cost, based on efficiency improvements alone.

One of the bigger benefits of automation, however, is the potential improvement in yield. The gains in consistency and process control that automation brings can deliver powerful material and labor savings on scrap. In a typical 980-nm pump pigtailing process, the cost of scrap and rework for each assembly can be as much as $150 in material and $50 in labor, which would generate more than $900,000 in annual savings, assuming a 15 percent yield improvement over 31,500 parts. The effect of yield improvement on the assembly cost per part, given various equipment costs, brings us to the conclusion that the assembly cost is most affected by yield (Figure 3). It is nearly insensitive to equipment-acquisition cost when running at high volume.

Figure 3. Assembly cost per part as a function of machine cost and yield shows yield — not equipment costs — as the driving factor.

Another important consideration for automation is volume (Figure 4). Here we can see the mix of machine vs. labor cost as a function of volume. Note that the labor component eventually dominates the cost equation because, at high volumes, the equipment contribution diminishes, but shift premiums add to the labor cost of running 24/7 operations.

Figure 4. A sensitivity analysis of cost vs. volume illustrates that machine costs per part decrease significantly with volume, whereas burdened labor rises as a result of shift premiums that must be paid to run a 24-hour operation. This example assumes a fully burdened labor cost of $50, including benefits and factory labor overhead.

Automation can result in significant savings. In most cases, we have demonstrated from 50 to 80 percent reduction in nonmaterial assembly costs. However, as we illustrated in Figure 1, automation can help reduce costs for only the labor, scrap and rework. It cannot inherently address the base material cost of the product, which represents the majority of the expense. Thus, other methods for driving down costs must be taken into consideration as well.

• Integration and design for manufacturability and automation: For this discussion, these two activities are combined because the first is really just one component of the other; i.e., it reduces part count. “Design for manufacturability and automation” techniques also encompass design rules and use of components that are inherently easy to see and to manipulate with machines; reduction in the number and variety of fasteners or assembly technologies; design for continuous-flow manufacturing; and self-registering components.

Examples of integration in the photonics world include combining multiple optical elements (or functions) onto one chip or substrate, monolithic integration of electronic components with the optical components in the wafer and multichip-module assembly.

Integration and design for manufacturability attack base costs in material by eliminating components and corresponding labor costs. They also can positively affect yield. Although the degree of cost reduction depends on the level of execution, this activity inherently offers high leverage, given that it can concurrently address costs in several areas.

• Standardization: Standardization of packaging components and processes could provide substantial cost relief, both in development as well as in material and processing. Aggregating large volumes of standard packages vs. having many custom packages in the market could reduce the cost of some components by as much as 50 percent. Furthermore, the reduction of design effort by using standardized components will save in development costs and time to market, as well as mitigate performance risks in the product.

• Offshore manufacture: The huge supply of cheap labor in Asia and Eastern Europe offers an attractive alternative to domestic manufacture. Looking at our cost profile, if you considered your labor as essentially “free,” you could possibly eliminate up to 25 or 30 percent of the production costs. But again, this brings with it the problems of training workers and supporting products in remote locations. Therefore, moving manufacturing offshore is not advisable for products that are new or for processes that are not well-characterized or -understood.

• Supply-chain management: An aggressive push on the material supply chain can yield a reduction of as much as 10 to 20 percent in a typical year. This is achievable by increasing volumes, leveraging more standardized designs, maturation of the manufacturing process and yield improvement — often via automation — in the vendor base.

This activity has tremendous leverage because it operates on the biggest cost element. This is one reason that subcontract manufacturers — i.e., manufacturing service providers — will likely become a solution of choice in more mature applications. They can aggregate larger volumes over multiple products and can leverage the supply chain in a more substantial way.

The pricing curve

Achieving sufficient cost reduction to keep up with market pricing pressures will require simultaneous attacks on several fronts. Automation can address labor savings plus yield-related cost recovery; offshore manufacturing can reduce labor costs (but is appropriate only for more mature products and processes); and supply-chain management can steadily work the material costs equation.

However, even successful combined efforts like these are unlikely to preserve sufficient profit margins. In forecasting cost-reduction efforts against an ongoing 25 percent sales price reduction over a four-year period (Figure 5), the following reasonable, yet fairly aggressive, cost-reduction actions are assumed:

Figure 5. Projected cost savings vs. price reductions shows eroding margins, even with aggressive cost cuts. Fundamental design changes, integration and standardization will be required to achieve next-generation targets.

• 10 percent per annum material savings through supply-chain management.

• 75 percent assembly labor savings over the forecast period through aggressive automation and continuous process improvements.

• 75 percent scrap and rework labor reduction over the projected period through automation and process yield improvements.

• 80 percent reduction in equipment cost per part due to increasing volumes and higher levels of automation.

The result is that, over the forecast period, costs decline by more than 50 percent, but margins still erode from nearly 40 to less than 10 percent.

A recent example of commercial success that supports this cost-reduction scenario is the Lucent Technologies (now Agere Systems) Laser 2000 program. First implemented in the late 1990s, this program entailed the radical redesign of a family of active components using design for manufacturability and automation, aggressive implementation of automation, flow manufacturing and supply-chain management. Although this program was implemented in the US (along with the commensurate higher labor rates), the company has reported a total manufacturing cost reduction of 50 to 75 percent over a three- to four-year span.

It is clear, however, that with continued market pricing pressure, fundamental product design changes also will be required to stay ahead of the game. Development and incorporation of substantial advancements in integration, extensive use of design for manufacturability and automation, and standardization of packaging components and manufacturing tools will be required to make photonic components economically viable for future-generation systems.

Fortunately, concerted efforts in each of these areas are already under way. And, in the meantime, the established techniques of automation, process improvement, offshore manufacture and supply-chain management will help keep component providers competitive.

Meet the author

Randy Heyler is vice president of business development for fiber optics at Newport Corp.’s Photonics Div. in Irvine, Calif.