Commercial Practices Shatter Defense Manufacturing Paradigm
by Dr. Douglas J. Carlson, M/A-COM Technology Solutions, Inc.

Introduction
Often, defense systems are characterized as too slow in development, delayed by technology which is not ready for production, unable to meet original specifications and too costly. Cost has often been cited as the dominant factor limiting the effective deployment of many major defense systems. As Secretary of Defense Robert Gates said in his recent Defense Budget Recommendation Statement, the DoD “must consistently demonstrate the commitment and leadership to stop programs that significantly exceed their budget or which spend limited tax dollars to buy more capability than that nation needs. Our conventional modernization goals should be tied to the actual and prospective capabilities of known future adversaries – not by what might be technologically feasible for a potential adversary given unlimited time and resources.” In this single statement, Secretary Gates identified two critical elements which plague many defense procurements: (1) high and/or uncontrolled cost resulting from delays and technology over-reaching and (2) system specifications that do not meet the actual requirement and continually creep, resulting in systems which do not meet the true need.

There is a stark contrast between the worlds of defense manufacturing and commercial manufacturing. In the commercial world, the mantra is speed and cost. If you are a viable supplier, it is assumed that you will meet specification – no one pays more for exceeding specification. The product development timeline defines the life or death of a new product. After all, Christmas only comes once a year. If you miss the product release window, the development effort is dead. Technology insertion in the commercial world is aggressive, yet the commercial environment does not forgive the supplier who does not meet requirements because of immature technology. Again, it is a given that you will meet the requirements. This is in stark contrast to the defense world, in which “Technology Maturation” programs are often implemented after the technology has failed to deliver. This results in years of program delays with inevitable cost growth.

In commercial programs, clear roadmaps for cost reduction must be established early on. Plus, these roadmaps must be achieved if a supplier is to remain viable in a competitive environment. As a consumer, when have you expected prices of goods to rise quarter after quarter? Imagine what the fate of a consumer supplier would be if, due to poor technology and manufacturing control, production yields suffered, resulting in continuous price increases to the customer. The company would go out of business. In the defense world, technology deficiencies, production inefficiencies and delays result in program increases – the customer pays more! As Secretary Gates indicates, these practices must stop.

M/A-COM Technology Solutions is demonstrating that the application of commercial manufacturing practices and a commercial supply culture can make a significant impact in the development and fielding of defense systems. Concepts such as Six Sigma, Design for Manufacturing, Design for Cost, and Lean Manufacturing are realities—not just slogans. These tools are critical elements to effective product development and manufacturing. They are tools to achieve the correct performance, on time, at the required cost. After product release, these tools help drive down cost to remain competitive.

In the case of electronics – specifically RF electronics – there is a dramatic difference in the evolution of technologies and manufacturing practices between the commercial and defense worlds. On the commercial side, RF systems have rapidly been driven to very high levels of complexity while maintaining a steady push toward miniaturization, energy efficiency, performance and reduced cost. The complexity of the RF front end of a cell phone has increased significantly over the past 15 years, accommodating many bands and communications protocols. Yet, RF electronics are consuming less board space, less power, and provide significantly more functionality through chip and board level integration. Defense systems over the same time period have, for the most part, maintained a reliance on low volume, semi-manual manufacturing of hybrid components. While there have been major advances in performance, these are accompanied by ever increasing cost. The application of commercial practices can help defense systems realize high performance, increased integration and reduced cost.

It is often asserted that consumer techniques cannot be applied to DoD systems for various environmental or perceived reliability issues. Yet, consumer manufacturing techniques are routinely used to produce components for very harsh applications in which human safety is a key factor. Products as diverse as those supporting our communications infrastructure to automotive electronics have demonstrated a high level of reliability while exploiting the most cost effective manufacturing approaches.

Examples

Packaging – Plastic vs. Ceramic
M/A-COM Technology Solutions is a major supplier of RF, microwave and millimeter wave components to both commercial and defense applications, and has been for nearly sixty years. Over this time, the company has gained a unique view of the various manufacturing approaches used for both commercial and defense applications. For example, a choice as simple as component packaging can have a profound effect on the total manufacturing cost. When considering packaging options for a MMIC, one can choose an industry standard, commercial, plastic PQFN style package targeted at consumer high volume manufacturing or a ceramic-metal hermetic package. This simple choice drives many factors, including production and test automation, options for the next level assembly and ultimately, cost.

Plastic PQFN packages have become the industry standard over the past several years for commercial RF components. In general, IC packaging is outsourced to very high volume production lines. These lines are highly automated, with rigorous production control resulting in very high production yields. The relatively high capital cost of the packaging line is amortized across the hundreds of millions of parts produced per month, resulting in very low per part cost. The packaging lines have well defined design rules coupled with fast production cycle time. This allows the designer to design in a well characterized environment and achieve rapid and controlled learning cycles in product development. The use of PQFN packaging also facilitated high volume, automated test, minimizing test costs. A further advantage is these package formats are amenable to high volume, automated board assembly operations, significantly simplifying the next level of assembly.

In contrast (See Table 1), metal-ceramic hermetic packages are produced on very low volume lines. Rather than hundreds of millions per week, volumes are thousands to tens of thousands per month – orders of magnitude lower! These low volume lines are usually manual or semi-manual as the volume cannot support a high level of automation. When and if a line is automated, the high capital cost is reflected in the packaging cost. In general, this packaging approach results in longer cycle times, slower learning cycles, increased production variability, weaker design rules and ultimately high part cost. Lower levels of control in production coupled with longer learning cycles stretches the design cycle and increases the design cost. Unlike commercial plastic packaging, the entire industrial base for this technology is small; hence, the industry will not be pushed forward, as commercial packaging is being driven by cell phones and microprocessors. In addition to the manual assembly, this package style leads to manual test and next level assembly – both driving increased cost.

Manufacturing Trade-offs
As a second example, consider a Transmit/Receive (T/R) module for use in Phased Array applications. A basic block diagram is presented in Figure 1. The T/R module consists of a number of RF functions: attenuators, phase shifters, gain blocks, low noise amplifier, power amplifier and switch or circulator. These blocks must be fed DC power and controlled in some manner to correctly manage the transmit or the receive signal. Additional complexities can be added to the T/R module to realize increased system capability.

Many approaches can be taken to implement this block diagram in hardware. We will consider a typical approach used for defense applications and an approach which exploits commercial manufacturing practices. In a conventional approach typical of defense applications, the T/R module would be constructed from discrete components using “chip and wire” technology. In this hybrid approach, a multi- layer ceramic substrate would be used as the integration media. As above, with ceramic packages, the design and manufacturing cycle for this substrate, commonly LTCC, is long. The RF components, which are often custom designed specifically for the application, are soldered to the ceramic substrate. Interconnections are achieved by wire bonding for the die to traces on the ceramic. In general, the designs are highly customized for each application with manufacturing lines running relatively low volumes. An unavoidable conflict between automation, cost and manufacturing control are encountered in this environment. This ceramic assembly would typically be contained in a hermetically sealed metal casing, adding yet more manual labor and cost. The form factor of the final assembly requires high cost custom testing fixtures.

Another approach can be taken. Examination of the Transmit (Tx) and Receive (Rx) sections of the block diagram reveals several features:

1. There is significant opportunity for integration of functionality both within each chain and between the Tx and Rx chains.

2. Each control component requires logical input to determine its correct state: switches, attenuators, phase shifters. This requires that designs must satisfy both RF and digital functionality.

In a commercial approach, partitioning of the block diagram is a key element to drive cost reduction through functional integration at the semiconductor level. Appropriate selection of functional boundaries will drive cost and performance. Partitioning will also either allow flexibility for future technology spirals without major redesign or prohibit them. On-chip integration is a clear path to lower cost when the volume of the opportunity is sufficient to warrant the custom design effort (Figure 2). Cost reduction through integration is parallel to the strategy followed by the silicon industry. In the case of digital processors, cost and performance is driven through integration and compaction reducing board component count, assembly complexity while dramatically increasing system functionality. The trade-off to consider is the cost and performance of custom design versus the use of COTS components.

In either case, there are several key elements which will be exploited in the commercial manufacturing approach:

• PQFN plastic packages – for low cost packaging and simplified, automated surface
  mount assembly of the T/R module

• Standard Printed Circuit Board (PCB) technology used as the substrate for the T/R module

• PCB production on high volume board lines

• T/R assembly on high volume, automated assembly lines

Using these manufacturing elements, the module can be designed high speed automated assembly and test (Figure 3). The approach described here for T/R modules is fundamentally similar to the approaches taken for the front end modules for cellular telephones as well as many other consumer applications.

M/A-COM Technology Solutions is actively engaged in demonstrating that commercial practices can help solve the cost, schedule and performance issues which are commonly encountered in defense programs. We believe that aggressive implementation of commercial manufacturing concepts will enable future systems to be developed and deployed rapidly, on time and with the performance the soldier requires.

M/A-COM Technology Solutions
www.macom.com
TXTLINX.COM113
Email this article to a friend!