Report on the Commercialization of Military & Space Electronics
The Commercialization of Military & Space Electronics workshop (CMSE) was held on Feb. 7-10, 1999 and is a successor to the Electronic Components for the Commercialization of Military and Space Systems workshops. This was the third in a series. The major subject matter of the conference is how to use commercial electronic parts in military and space systems where military or special grade parts were used in the past. Frequently, the commercial electronic parts are not certified to operate over the temperature ranges which are required by military systems. In other cases, parts must be used in a radiation environment or in other situations which are not covered by the manufacturer’s data sheet.
In general, people are using Commercial Off The Shelf (COTS) parts in "high reliability" military and space systems and are generally ignoring component manufacturer’s proscriptions against using parts beyond their data sheet ranges. There were many presentations on how to do this ("Uprating" and/or "Upscreening") and much discussion, both in panel discussions and in the halls.
The workshop notes are available from the Components Technology Institute (256-536-1304). They contain all the viewgraphs used in the presentations plus a few papers amplifying some of these presentations.
Review of selected papers
Since my own interests at this time are devoted primarily to the qualification of COTS for use in high reliability applications, I have chosen some papers of special interest in this area for short reviews.
J. Kinnison and co-workers from the Johns Hopkins Applied Physics Lab and US Army AMCOM presented a paper titled Results from a Plastic Encapsulated Microelectronics (PEM) Qualification Study in which they discussed qualification of plastic parts for military applications. In this paper, the parts are actually identified by manufacturer and part number, an unusual but very desirable feature. The investigators attempted to use C-mode Scanning Acoustic Microscopy (CSAM) to detect or quantify package delaminations which could be correlated with functional "failures" after Highly Accelerated Stress Test (HAST) or Temperature Cycling (TC). They reported that the correlation was "unclear". This is a general result found by many investigators. The CSAM is a valuable tool and provides insight into PEM package quality but failures in HAST and TC can occur with or without delaminations. These investigators also concluded that each individual "lot" of PEM components must be tested because there can be significant lot-to-lot variability. This conclusion is also supported by other investigators and provides support for a "lifetime buy" strategy, particularly in programs where small numbers of parts will be needed.
Tom Davidson from NSWC Crane presented a paper titled PEM Up-Rating Evaluation. He made extensive use of CSAM in his studies and suggests that both CSAM and electrical measurement should be done in "Uprating" environmental tests. As expected, the major changes in CSAM images were found after part. pre-conditioning. No further delaminations or other changes were seen in the course of the environmental tests; HAST, TC, thermal shock,..
Bill Glover of the US Army AMCOM presented a paper titled, Evaluation of Plastic Encapsulated Microelectronics for Missile Applications. A major AMCOM missile requirement is high reliability after long term storage. Bill described a very aggressive environmental test program for PEM involving sequential HAST, TC, salt exposure followed by more HAST and TC. Such sequential testing is not prototypic or characteristic of anything which could happen to the PEMs in real life. For example, the 130°C/85% RH 100h HAST will certainly weaken the Au-Al wire bonds, as shown by many investigators. Subsequent part thermal cycling (100 cycles, -55 to +125°C) could well produced a wire bond failure. However, if parts can pass such a demanding test, they demonstrate extreme ruggedness with regard to environmental stressing. In agreement with many investigators, Bill does not feel that selection of PEMs for use in demanding applications can be based on vendor reliability data. He stated that the "amount of testing" may be determined by the availability and quality of these vendor data but that some testing will always be necessary. In the future they will try to correlate experimental results with reported vendor data.
Mike Sandor and co-workers from the Jet Propulsion Laboratory presented a paper titled Commercial Off-The-Shelf Program, Issues and Results of Upscreening COTS for NASA Flight Hardware. They put their qualification effort into Destructive Physical Analysis (DPA), TC, and CSAM. Moisture performance of PEM was not a critical issue in their satellite application. They felt that their test program was effective in reducing mission risk and felt that the value added by the qualification in risk reduction exceeded the qualification costs. In using CSAM they found an interesting condition in which there was apparent die surface delamination on all parts of one type. However, the apparent delamination was produced by the use of a die coating which obviously had a sound velocity less that that of the mold compound. Hence the reflection at the mold compound – die interface produced a phase inversion similar to that produced by a delamination. This example indicates that CSAM data requires some interpretation and that it is not a "turn key" technique for assessing package quality.
In another paper from the Johns Hopkins APL, Andrew Moor and Kim Cooper analyzed the costs of using commercial vs. military parts for an application. The paper is titled, Costs Affecting the Choices Between Using Commercial & Military Parts in Space Applications. This work is of interest because there is significant controversy about the use of commercial parts to save money. While it is certainly true that the cost of a purchase order for a group of commercial parts will be less than that for an equivalent group of military parts, the costs for qualification, storage, and assembly of the commercial parts may well exceed the purchase order savings. In this paper, Moor and Cooper attempted to quantify the commercial to military part cost ratios of certain cost components. For example, they found that the labor costs for commercial part selection exceeded those for military part selection by a factor of 8. This occurs because the selection of commercial parts requires evaluation of the vendor and vendor reliability data while military part selection requires only reading a catalog and ordering the desired parts. The cost ratio for the labor in screening parts was reported as 3.3. While military parts are burned-in and commercial PEMs are not, the qualification costs of running environmental tests on PEM far exceed the saved burn-in cost. Their conclusion was that COTS should not be used just to save money. If mission enabling military parts are available, their use will result in the cheapest system. However, they felt that in many cases, use of COTS parts would result in lower system costs because of the non-availability of military parts with the correct function or form factor.
Mark R Troy, a lawyer who specializes in government contracts, discussed how Defense Contractors using COTS may be open to civil liability claims. In the good old days the components were proscriptively designated by DoD so the contractor could claim they had no choice in picking the parts.
Summary
The few papers described above represent just a small sample of the many excellent works presented. The conference was well organized and appeared to be well received by the participants. Although not much is new or original in this field, there is much to be gained by sharing experiences on COTS and PEMs qualification and use in high reliability systems. In addition, the evolving body of knowledge and experience may well help in the development of standard test methods. Also, COTS vendors may improve their reliability testing and documentation in order to capture more of the admittedly small but influential military and high reliability market.
--submitted by James Sweet, Sandia National Labs