GEM : An innovative solution to the DMS Problem

Lloyd S. Peters

Sr. Program Manager
Computer Science Laboratory
SRI International

1. Introduction

This country continues to face a serious military logistics problem with respect to Diminishing Manufacturing Sources and Material Shortages, known as DMSMS or DMS for short. The focus of this paper is on microcircuits and the part the expanding technology of Generalized Emulation Microcircuits (GEM) is playing to reduce the magnitude of that problem.

2. Background

In the rapidly evolving microelectronics industry, emphasis is placed on keen competition in both design and manufacturing. This competition results in a very rapid pace in the development, implementation, and replacement of both microcircuit processes and the products produced. This strategy is suited to commercial markets where the first to market can enjoy significant financial reward. However, it causes significant problems for military systems where the design cycle is typically longer than the time the microcircuit components are in production. Making the problem even worse is the military requirements for a long-term supply of microcircuits for follow-on production or logistics support. The limited production quantity and inherent complexity of military systems make redesign (to work around nonavailable microcircuits) or upgrade costly in both dollars and time. This discontinuance or diminishing manufacturing sources results in a significant problem that in recent years has become a major problem for the Department of Defense, NASA, and other government organizations.

While the DMS problem exists in many industrial commodities used by DoD, by our estimates it is most severe for microcircuits and expected to increase in severity in the future. It is estimated that the DMS problem exists for 40 to 60% of the microcircuits used in fielded military systems. The problem is not limited to fielded systems, as there is evidence of DMS problems with systems under development before the design and acceptance test cycle is completed. The original manufacture of DMS microcircuits can date as far back as twenty-five years to as recent as four to five years ago. The bulk of parts where the need is currently critical can be characterized as medium- to large-scale integrated (MSI/LSI) circuits, with a few parts in the very-large-scale integrated circuits (VLSI) category. Projections based upon manufacturing discontinuance models indicate that the demand for the MSI/LSI parts will continue to increase with the dramatic increase in demand for VLSI parts in the next three to five years. The majority of demand (that cannot be satisfied) is for digital microcircuits. There are a few linear and power microcircuits for which there are no sources, and projections indicate a significant increase in DMS for selected types of these devices within the next five years.

The absence of sources for microcircuits needed for manufacture or repair of military systems places undue pressure on managers responsible for operational readiness. The response to this pressure is typically to authorize redesign or system upgrades to provide an interim solution to the microcircuit nonavailability problem. In the best of circumstances the redesign will resolve the problem for the remaining operational life of the specific system. However, it is typical to find that the same microcircuits (for which the redesign was authorized) are needed by other military systems resulting in the proliferation of redesign activities and what we refer to as the obsolescence cycle. Since the life of a microcircuit manufacturing source is typically only a small fraction of the service life of the military system it is not atypical to find that more than a single redesign is required during operational life of a military system.

The GEM research, development, and validation program was developed to provide an alternative microcircuit source to assist in resolving the above described problem.

3. GEM Emulation Technology

GEM was developed during a three-year period and has been producing emulated replacement microcircuits for validation purposes (including testing in selected military systems) for more than three years. The objective of the GEM program is to define, develop, and demonstrate a generic emulation system that makes the best use of modern technologies to economically produce microcircuits that are form, fit, and function (FFF) equivalent to devices originally produced by obsolete technologies. The GEM system provides a new technology option suited to the support of weapon systems throughout the 1990s and beyond.

The developed technologies include a BiCMOS process, specification approach, design environment and methodology, and fabrication system: all tailored for generic emulation.

The GEM BiCMOS process was developed to support the requirement that the technology could be easily transferable to industry to ensure long-term competitive sources for emulation in support of Government needs. The GEM BiCMOS process design permits standardizing on a single mainstream process that can produce FFF digital IC devices that might otherwise require multiple (more than 10) processes and suites of fabrication equipment.

The specification approach provides for reconstructing design and manufacturing specifications through characterization testing that complements all sources of available technical information. Reusability is incorporated in the testing environment by integrating the test criteria with the design environment and using standardized procedures on automated test equipment.

The design environment makes use of commercially available Mentor Graphics Corporation computer-aided engineering (CAE) tools, GEM array designs, and a GEM-developed design library to support rapid production of FFF ICs at minimum costs. The use of an application-specific IC (ASIC) approach, together with the GEM design library, provides the capability to maximize use of previously developed designs and design building blocks to minimize the time, effort, and costs of for each additional IC type.

The fabrication system makes use of the David Sarnoff Research Center's IC Center foundry for developing the GEM BiCMOS process, enhancements, and production of GEM wafers for inventory and on-demand personalization.

The integration of the major GEM system blocks (respecification, fabrication, and testing) resulted in the complete emulation system shown below.

The current GEM capability provides immediate benefits by making available microcircuits that would otherwise be unavailable. In addition, the existence of a GEM system provides value in areas that are less obvious when viewed from the parts sources perspective.

4. GEM Operations

The GEM program continues development on two tracks (1) the advancement of generic microcircuit emulation technology for the purpose of increasing coverage and reducing costs in pursuit of the most complete and least-cost logistics support for electronic systems, and (2) designing, producing, and testing emulated microcircuits needed to support production and maintenance of military systems to validate the efficacy of the current emulation capability.

Applications experience to date has demonstrated the GEM capacity to produce needed microcircuits, on demand, while saving the Government millions of dollars. The experience has also shown that GEM is affordable technology with clear, measurable cost savings. The GEM program has been producing emulated ICs since 1989. Two examples of how GEM has provided improved logistics support to our military are described here.

The AN/UYK-44 computer is the Navy standard computer for shipboard use. When a follow-on production order was made for additional computers, it was determined that one micro circuit (Generic Part No. 25LS2513) used in four different modules was out of stock, with no production sources available. The GEM replacement was specified, designed, and produced within eight weeks. Sample emulated parts were tested at the component, printed wiring board, and system level over the military temperature envelope. The tests were completed rapidly, and there were no failures. The parts are currently being manufactured and provided for production of the AN/UYK-44. The availability of the emulated parts will produce a savings of $1.2 million, a fifteen-to-one cost reduction. The demonstrated cost savings of the initial version (in a flat pack) of this emulated part has resulted in a request to produce a dual in-line (DIP) packaged version for use in supporting other military equipment.

The second example is the APG-63 radar, which is used in some models of the F-15 aircraft and which was experiencing significant IC nonavailability problems. In response to a request from the Air Force, a study was conducted to define how GEM might produce microcircuits for use in support of the F-15. The study concluded that GEM could provide 86% of the needed ICs for the APG-63 radar and had the potential of saving hundreds of millions of dollars if a major system upgrade could be avoided.

The GEM program produced two demonstration parts that were thoroughly tested by the Warner Robins Air Logistics Command (ALC) and shown to perform correctly in a number of different applications. Currently, the GEM system is being used under a contracted task order arrangement to emulate 14 different microcircuits that are needed by Warner-Robins ALC for long-term support of the F-15 aircraft. These parts will be tested to demonstrate that they fully satisfy the FFF criteria and will then be qualified and integrated into the microcircuit supply channel.

The success of the GEM program in demanding situations has increased the interest by original equipment manufacturers (OEMs) and replacement microcircuits are now being emulated for the many of the major aerospace concerns. One example particularly noteworthy is an effort where a GEM part will be designed and built in cooperation with Rockwell - Space Division to assess the feasibility of using GEM to relieve the space shuttle/orbiter DMS situations.

5. Summary of GEM Potential within the DoD

The GEM technology has progressed to the point that it can provide high-quality, low-cost FFF microelectronics that will be available to support weapon systems throughout their operational life cycle. The nonrecurring engineering (NRE) cost of the GEM option is approximately 20% of the nonrecurring engineering (NRE) cost of single circuit board redesign. The systematic use of GEM could save the Government millions of dollars over the next 2 to 5 years, since more than 40,000 ICs are extremely vulnerable to DMS.

Since GEM replacement parts adhere to the strict FFF criteria, the parts can be used as plug-in replacements without incurring the typical documentation difficulties normally involved when alternative parts are substituted. GEM is one of the few DoD programs with significant growth potential during the period of declining DoD budgets. The need for replacement microcircuits is being increased by service life extensions of military equipment. In addition, policy directives that will result in modernizing DoD's inventory practices will accelerate the long-term need for GEM, a technologically superior solution.

6. Future Direction

What is the Future of GEM? In a word, we feel the future of GEM is bright. The system is operational. The system is producing FFF microcircuits of military quality in time for and in the quantities needed for logistics support. GEM is a proven, up-and-running system that has met every reasonable challenge presented to it. There is little doubt that GEM will be around for the long term. As a consequence, we are concentrating on reducing NRE costs and are constantly improving/expanding the technology to stay ahead of evolving needs. Finally, we are doing everything we can to make the GEM system known to a ever-increasing spectrum of potential users and to transfer the technology to the industrial base.

Last updated 04/21/95