GEM DMS Paper

The Potential to Reduce the DMS Impact on the Space Shuttle Orbiter

Ronald Datilio
Logistics Operations Project Manager,
Rockwell Space Systems Division

Lloyd S. Peters
Sr. Program Manager, SRI International
1. Introduction
The impact of diminishing manufacturing sources (DMS) for microcircuits is not an exclusive DoD problem. The quality of microcircuits built to DoD M38510 specifications attracted NASA and others to specify these high-reliability devices in the design of critical systems. The reduction in space exploration budgets and the planned service-life extension of the space shuttle/orbiter increased the importance of a proactive approach to addressing DMS issues. The appendix to this paper provides a perspective on the potential impact that diminishing sources may have with regard to a set of microcircuits designed into the space shuttle/orbiter.
The Office of the Secretary of Defense advised NASA of the DoD-sponsored Generalized Emulation Microcircuit (GEM) technology development program that could provide an economical source for otherwise nonprocurable microcircuits. In recognition of this situation and the opportunity to participate in the GEM demonstration/validation (DEM/VAL) activities, NASA sponsored a project to evaluate emulation as a possible source for no-longer-procurable (nonprocurable) microcircuits. NASA requested that Rockwell (the orbiter prime contractor) perform the pilot project under the Space Shuttle Safety and Obsolescence Upgrade (S&OU) program. The objective of the pilot effort was to evaluate the GEM technology, including an assessment of the potential benefits of emulation in substantially decreasing the impact of microcircuit obsolescence. The Rockwell assessment included a thorough evaluation in accordance with its standard policy.
Manned space systems, such as the space shuttle/orbiter, require large capital investment and are designed for long life with systems reliability being a paramount concern. In this context, any replacement of microcircuit components requires a thorough evaluation of the replacement components to ensure that they meet or exceed these reliability requirements. To adhere to this overriding purpose it was essential that the pilot program be conducted with a high level of collaboration among NASA, Rockwell Space Systems Division, and the GEM program team. The level and extent of the cooperation achieved was exceptional and was evident at all technical and administrative levels. In the opinion of the authors, the exhibited teamwork was the key element in the success of the effort.
2. Background of the Space Shuttle/Orbiter Preferred Parts Strategy
In building the space shuttle/orbiter NASA/Rockwell used the strategy expressed in the report "Electrical, Electronic, and Electromechanical Orbiter Project Parts List," 1 February, 1988. The strategy was to select the readily available microcircuit standard products of the highest obtainable quality. This led to the following rules:

Utilize parts that are adequately controlled by existing specifications and, preferably, manufactured in the United States.
Utilize qualified parts with proper quality conformance inspections.
Utilize established reliability parts that have demonstrated successful application.
Reduce the part types used and the number of specifications or drawings.
Standardize burn-in and screening requirements.
Standardize stress derating for critical parameters.
Standardize traceability requirements.
Utilize Destructive Physical Analysis (DPA) for configuration, process, and quality control verification to assure consistency of parts.

A thorough review of these rules reveals that they are logical and serve a very important purpose of standardization (of significant importance in good logistics practice), while ensuring good design practice. However, these objectives could not predict the changes that were forthcoming, including:

The explosive growth of the microcircuit industry and the wholesale discontinuance of microcircuits
The advent of budget-induced restrictions in the upgrades and redesign opportunities
The service life extension of the space shuttle/orbiter

The results of the these three external influences is that many of the microcircuits that were readily available when designated and designed into the space shuttle/orbiter are either not available or have a single source and are vulnerable to becoming nonprocurable. This situation highlights the importance of identifying a long-term, cost-effective source for microcircuits in support of the space shuttle/orbiter.
2.1. Microcircuit DMS Problem
The DMS problem leading to nonprocureability is most common in microcircuits. The Defense Logistics Agency (DLA) manages well over 4 million different items, of which microcircuits number approximately 115 thousand. While the total number of microcircuits represent less than 3% of the total managed items, they account for more than 55% of the nonprocurable item actions. Microcircuits are at least 41 times more likely to develop a DMS problem than the other DLA managed items categories.
Microcircuit obsolescence occurs when new technologies, processes, and products displace the sales volume of older devices, hence rendering the manufacturing of the older products too costly with the in-place fabrication system.
The defense market -- once the mainstay of the technology push to the semiconductor industry -- is no longer a major player and has limited influence in preventing such obsolescence. Current estimates are that the DoD share of the U.S. microcircuit market is less than 5%, and of the world market perhaps less than 1%.
Life cycles of microcircuits are significantly shorter than life cycles of defense and other high-reliability systems. During the formative years of the semiconductor industry (more than 30 years ago) typical life cycles of mainstream microcircuit processes and products were 15 to 25 years; current life cycles are 3 to 7 years.
During the 1960-1970 period the fledgling semiconductor industry with Government influence promoted the establishment of standard products that were typically produced by multiple sources. Currently, industry leaders are concentrating on establishing niche positions with the discontinuance of standard products in favor of high-use specialty products (e.g., microprocessors, memory devices, programmable logic devices, special-purpose processors) and application/customer-specific devices. The life cycle of a niche product line can be as short as 18 months.
2.2. Dealing with Obsolescence
Most prime contractors of large Government-owned systems are now painfully aware of the microcircuit obsolescence problem. Increased attention is being given to microcircuit availability concerns because of the double impact of the discontinuance of sources (nonprocurable components) and service life extension for major Government systems. At the same time, the aerospace industry is being faced with budget constraints that compound this situation.
In the past, the standard industry approach was to plan the future needs for microcircuit components based on estimates and expected equipment redesign for upgrade purposes. This standard approach emphasized inventory management (dependence on manufacturing sources or stocking of parts that are expected to be needed for planned production and support, with the fallback of exercising life- of-type buy options) and the redesign option.
Why is this proven approach not providing the desired results? First, both the scope and rate at which microcircuit products are being discontinued are accelerating. Second, the budgets for defense and space systems are being reduced. Under less constrained budgets, redesign was regularly performed and any obsolete components were removed from the design during the next planned upgrade cycle. With the budget constraints, redesign is an infrequent activity, and available funds are being used for essential upgrades (to add new capability) or to fix identified reliability problems. Funds are not available for redesigns necessitated by parts obsolescence.
The inventory/redesign approach suffers from market forces that increase the price for short-supply replacement components, and the budget-imposed reduction in redesign frequency precludes economy-of-scale scheduling to reduce the cost per unit.
3. Emulation as the Newest Technology
Emulation is a technology that in the past received limited application to the microcircuit obsolescence problem. In most attempts, emulation approaches suffered from trying to use off-the-shelf technology that at best resulted in limited applicability, was expensive, and lacked long-term availability.

The GEM program was initiated for the purpose of developing and demonstrating a generic form of emulation specifically tailored to addressing microcircuit obsolescence. The developed GEM technology offers a new dimension for dealing with microcircuit DMS problems. Although not the sole solution to this problem, GEM technology does offer an approach that can provide order-of-magnitude reductions in the cost and difficulty of supplying microcircuits for logistics support of electronic systems. It complements the use of existing inventories of nonprocurable microcircuits and design upgrades.
How does GEM technology achieve such leverage? In answering that question a brief description of GEM is in order.
GEM was developed by a team of researchers from the DLA, SRI, and The David Sarnoff Research Center. This team had broad knowledge of both the microcircuit process and product technologies originally used to produce the devices that were either nonprocurable or approaching a nonprocurable status. The objective of GEM was to develop a technology to emulate microcircuit devices (some originally produced more than thirty years ago) by using the latest state-of-the-art design and fabrication tools.
The challenge for GEM development was to produce a generic emulation capability emphasizing innovation and using a single process technology and unified approach that could reproduce devices originally built with multiple technologies. It is through this invention of a complete system architecture built upon specification, design, fabrication process, and testing technologies that GEM (the generic approach) provides a unique capability. Some of GEM's features are

Flexibility to tailor the electrical characteristics of emulated microcircuits to match function and I/O characteristics of devices previously built with process technologies covering the past thirty years
A unified approach that permits the capture of proven design elements, designs, test programs, and procedures that can be reused for a significant portion of the nonprocurable microcircuits
The use of fabrication mosaics that permit production of multiple different microcircuits (function, technology, and signaling characteristics) on a single wafer and significantly reduce the minimum economical order quantity

4. Rockwell Experience with GEM Technology
The scope of the pilot effort included the following specific tasks:

Selection of a candidate nonavailable part for emulation
Evaluation of emulation approaches and selection of the best approach
Development of the pilot requirements and establishment of schedules
Design, fabrication, and piece-part testing of the part selected for emulation
Definition of support system requirements to accommodate an emulated microcircuit device
Insertion testing of the emulated part to validate its correct performance in the circuit board and line replaceable unit (LRU).
Definition of the requirements to incorporate successful results into program documentation The results of the collaborative efforts of NASA, DLA, Rockwell, and the GEM team were as follows:

Rockwell, in consultation with NASA, selected the 54L20W as a was nonprocurable part that was designed into four different space shuttle/orbiter subsystems.
The technical data package of all available documentation for the 54L20W was compiled, including the part specification, application circuit drawings, and typical part test results. The specifications for the emulated version of this part were prepared, reviewed, and approved.
A Rockwell audit of the Sarnoff manufacturing facility was performed. An assessment report was prepared that approved the Sarnoff manufacturing process as suitable for producing microcircuits for manned space flight applications.
The schedule for the emulation efforts was established as follows: Design completion 1/94 Wafer fabrication 3/94 Package & 5004 screen 4/94 ESD protection testing 5/94 Insertion test (C&W system) 6/94 Radiation hardness tests 7/94.
Establishing the support system requirements required resolution of a number of issues before emulation technology could be authorized for use in manned space flight systems:

How to authorize the installation of an emulated microcircuit that carries a part number different from that on the Bill of Materials.
How to ensure that an LRU containing emulated microcircuits is certified
How to track repair-driven attrition use of emulated microcircuits.

Desk Instructions that address the above three issues and are entitled "How to Authorize Installation/Use and Incorporate GEM Microcircuits into Program Documentation" are currently under review within NASA prior to approval and implementation.

Another issue of concern to both Rockwell and NASA was the quality of microcircuits used in manned space flight. To address this issue, both the facilities used to manufacture the GEM microcircuits and the procedures for qualifying specific microcircuits were evaluated:

The Sarnoff manufacturing facility was assessed and approved as stated above.
A quality test requirements document was developed that delineates specific testing required to verify quality conformance. The emulated 54L20W was subjected to the required testing with the exception of selected Mil-Std-883D Quality Conformance Inspections that are scheduled for completion late in 1995.
Testing and evaluation of the emulated 54L20W device produced the following results:

Emulated 54L20W part with functional performance equivalent to of the original
ESD device protection of the GEM part established by test as superior to that of original parts
Emulated part exhibiting tighter electrical characteristics than the original part
Radiation environment tests (being performed per NASA standards)
Emulated part passing all system insertion test specifications (functional, temperature, vibration, ATP)
Emulated part installed in Caution & Warning LRU and returned to the Shuttle Avionics Integration Laboratory (SAIL) for long-term reliability tests.

The form, fit, and function (FFF) emulated 54L20W microcircuit satisfied all specifications.
Tests were made on emulated 54L20 parts in both flat-pack and DIP packages planned for insertion testing in the PCM, MMU, and IMU sub-systems.

Thus, FFF 54L20W microcircuits satisfied all specifications, and the pilot program experiment has established that the GEM technology provides a cost-effective long-term source for microcircuits that would otherwise be nonprocurable. The pilot program for piece parts has met all of its objectives and will soon be in place to provide microcircuits as needed for the space shuttle/orbiter. The emulation procedures have been demonstrated and are well on the way to implementation. A review of GEM technology has provided evidence that it could be adopted to provide a replacement source for nonprocurable hybrids and circuit boards.
5. Conclusions
The first space shuttle/orbiter GEM emulated microcircuit developed by SRI International in conjunction with its team member and subsidiary the David Sarnoff Research Center, was delivered to Rockwell Space Systems Division in April 1994. The part insertion testing, followed by technical evaluation, was successfully completed at the end of June 1994.
The Piece Part Emulation Pilot Program has demonstrated that emulation using the DLA sponsored GEM technology is suitable for space shuttle/orbiter microcircuit replacements. This emulation technology has demonstrated tighter margins of electrical characteristics over temperature and superior ESD protection with a potential for safety and reliability enhancement through technology insertion.
One single best solution to the nonprocurable microcircuits situation is not universally applicable. Emulation technology can provide significant cost and availability benefits as a component of the logistics support equation in concord with managed inventory and design upgrades strategies.
The significant achievement for the space/shuttle orbiter program illustrates the close teamwork that was experienced within Rockwell, with IC manufacturers, and among Government agencies focused on reducing costs and extending the service life the space shuttle program.
Other findings for the performance of the emulation pilot program are that:

The problem of obsolescence and the consequence of nonprocurable microcircuits is real. The need for microcircuits to support follow-on manufacturing and logistics support is being significantly impacted.
The standard approaches of inventory management and redesign are inadequate as the only solutions in the context of service life extension of major systems.
Emulation as implemented by the GEM program is a technology that can improve logistics support. The insertion of GEM technology adds a dimension that will permit targeting limited fiscal resources to priority upgrades and will help manage escalating microcircuit inventory costs.

Appendix: Perspective on DMS Microcircuits Used on the Space Shuttle/ Orbiter
To obtain a perspective of the potential impact of microcircuit DMS on the space shuttle/orbiter, a technical assessment was performed on a selected subset of the incorporated microcircuits. The parts subjected to the technical assessment represent only a portion of the microcircuits needed for the space shuttle/orbiter.
The assessment of the types and probable availability of microcircuits was designated in the Rockwell International document, entitled: "Electrical, Electronic, and Electromechanical Orbiter Project Parts List", dated February 1988. This document is referred to as the OPPL.
The results of the assessment of type and probable availability of the orbiter microcircuits provide a good perspective on the DMS situation that currently exists regarding space shuttle/orbiter standard microcircuit parts availability. The OPPL analysis concludes that

All designated parts are specified by reference to M38510/ numbers.
There are 318 different functional part types: 254 digital devices and 64 linear/analog devices.
Digital devices represent 80% of the total and linear devices represent 20% of the total.
A majority of all the cataloged manufacturing sources for the OPPL microcircuit device configurations have been discontinued. For the digital microcircuits this amounts to 76% and for the linear microcircuits 46%.
For the OPPL microcircuit devices 57% of the digital devices have no source or a single manufacturing source, and 67% of the analog devices have no source or a single manufacturing source.

It should be recognized that the microcircuits assessed represent a recent (1988) designation of microcircuits for use in the Shuttle/Orbiter and therefore are probably more readily available than parts used in previously built subsystems.
A comparison of this list of parts to the current GEM capability established that GEM could provide sources for most (98%) of the DMS digital devices and more than 10% of the analog devices.
GEM HOME

Last updated 04/21/95
iskender@csl.sri.com

Send any questions or problems regarding this service to iskender@csl.sri.com.