By Dan Friedlander - Retired following 44 years in component engineering
Introduction: Recently, a COTS in Space workshop ACCEDE 2019 (Assessment of Commercial Components Enabling Disruptive space Electronics), organized by Alter technology, has been held in Seville Spain on 6th thru 8th November 2019. The workshop has been attended by 250 participants from 25 countries, from all the world (Europe, USA, Canada, Japan, China, Argentina, Brazil etc.). I was honored to give a keynote speech. The speech was somehow unconventional, presenting the need for change versus resistance to change, due to global developments.
For me, just to stir things up seems a great reward in itself.
It seems that a huge progress has been made in the domain of recognizing the viability of using EEE COTS components in space applications. However, the process of creation of new official methodology/policy, to meet the present and future challenges, is too slow. Meanwhile, a large variety of local methodologies have been implemented, bypassing the official outdated ones. The details of these local initiatives are not disclosed by the relevant parties due to competition reasons. Considering the above facts, it is not understood (at least by me) how a prime contractor is supposed to deal with a large variety of policies of the relevant subunits in one system! There is not a proper document to refer to. In the above environment how the ESA ECSS-Q-ST-60-13C (21 October 2013) "basic approach" can be materialized? Quote from the ESA spec: " - The supplier defines a component control plan to implement those requirements into a system which enables, for instance, to control the selection, approval, procurement, handling in a schedule compatible with his requirements, and in a cost-efficient way.
- The supplier ensures that the applicable parts requirements are passed down to lower level suppliers and ensure that they are compliant to these parts requirements."
This article deals with some issues discussed during the above mentioned workshop. ESA ECSS-Q-ST-60-13C (21 October 2013) - Space product assurance commercial electrical, electronic and electromechanical (EEE) components. It has been stated that the ESA ECSS-Q-ST-60-13C (21 October 2013) will be soon upissued. However, it is worth the time to clarify some points.
The ESA specification covers only integrated circuits, not addressing other EEE COTS component types. That creates confusion.
Quote from ESA spec: " This standard applies only to commercial components - as defined in its scope - which meet defined technical parameters that are on the system application level demonstrated to be unachievable with existing space components or only achievable with qualitative and quantitative penalties."
This requirement penalizes the use of EEE COTS components versus use of existing space components. COTS are considered as last resort! Of course, practically, the user will find the way to bypass the above requirement. In view of statements made in the workshop on the issue of the ESA spec not mandating requirements, the following quote clarifies the intention:
Quote from ESA spec para. 3.5):
" a. The word “shall” is used in this Standard to express requirements. All the requirements are expressed with the word “shall”.
b. The word “should” is used in this Standard to express recommendations. All the recommendations are expressed with the word “should”.
NOTE It is expected that, during tailoring, recommendations in this document are either converted into requirements or tailored out."
Quote from ESA spec (para. 4.3.3d ): " For commercial parts, screening tests shall be performed in accordance with Table 4-2." That means, interalia, that the heavy burden of postprocurement 100% screening (moved for EEE COTS from the component manufacturer to the user) is a mandatory requirement in accordance with the ESA spec. The word "shall" is extensively used across the ESA spec. NASA PEM-INST-001 (June 2003) - Instructions for Plastic Encapsulated Microcircuit (PEM) Selection, Screening, and Qualification.
This a Goddard Space Flight Center (GSFC) Technical Publication widely referred to, dealing with plastic encapsulated microcircuits (PEMs).
Quote from NASA spec: "Due to the major differences in design and construction, the standard test practices used to ensure that military devices are robust and have high reliability often cannot be applied to PEMs that have a smaller operating temperature range and are typically more fragil and susceptible to moisture absorption. In contrast, high-reliability military microcircuits usually utilize large, robust, high temperature packages that are hermetically sealed."
Standard test practices can be applied to PEMs, except the irrelevant ones (e.g. hermeticity, PIND). The plastic packages are smaller, more robust, less frail, more suitable at 2nd level connection (temperature variations). The moisture absorption is manageable (no need for hermeticity). The military operating temperature range is narrower in space applications. The widely use of "high-reliability military microcircuits" implies that COTS cannot be high-reliability components!
Quote from NASA spec: "Unlike the military high-reliability system, users of PEMs have little visibility into commercial manufacturers’ proprietary design, materials, die traceability, and production processes, and procedures. There is no central authority that monitors PEM commercial product for quality, and there are no controls in place that can be imposed across all commercial manufacturers to provide confidence to high-reliability users that a common acceptable level of quality exists for all PEMs manufacturers. Consequently, there is no guaranteed control over the type of reliability that is built into commercial product, and there is no guarantee that different lots from the same manufacturer are equally acceptable. And regarding application, there is no guarantee that commercial products intended for use in benign environments will provide acceptable performance and reliability in harsh space environments."
The use of the wording "guarantee" is inappropriate. No central monitoring authority can guarantee the level of reliability, quality and uniformity of any EEE Components. Refer to the history of space/military components' related failures and alerts.
The goal is not to reach the same EEE COTS level of reliability offered by space/military microcircuits. The goal is to reach an acceptable level of reliability, within the technical, availability and financial constraints. The EEE COTS components have been used successfully in commercial and military harsh environments. The space is not a more severe environment than encountered in many military applications, except the radiation. The selected COTS shall meet the radiation requirements, like any selected space/military components. The radiation withstanding capability of an EEE Component is established by the die and not by the type of package (hermetic, nonhermetic) or the quality level (space, military, commercial).
Quote from NASA spec: "Product assurance methodology of most PEM manufacturers is based on the philosophy that the reliability must be designed or built into the manufacturing process rather than achieved by 100 percent testing of products. According to this methodology, the emphasis is on increasing the yield and reducing the likelihood of defective parts production by tight process control rather than on detection of failures during electrical testing."
This philosophy works well for high volume production, as practiced in the commercial industry. The military/space traditional methodology is based on testing, being a low volume production. The strict regime of heavy testing rules led to the huge technological gap between the commercial world and the space/military one. Quote from NASA spec: "The qualification and screening processes contained in this document are intended to detect poor-quality lots and screen out early random failures from use in space flight hardware. However, since it cannot be guaranteed that quality was designed and built into PEMs that are appropriate for space applications, users cannot screen in quality that may not exist. It must be understood that due to the variety of materials, processes, and technologies used to design and produce PEMs, this test process may not accelerate and detect all failure mechanisms. While the tests herein will increase user confidence that PEMs with otherwise unknown reliability can be used in space environments, such testing may not guarantee the same level of reliability offered by military microcircuits. PEMs should only be used where due to performance needs there are no alternatives in the military high-reliability market, and projects are willing to accept higher risk."
According to this document in case of EEE COTS components usage in space applications the screening burden is moved from the component manufacturer to the user. Regardless of who is performing the screening, the technical scope of the testing stays as is. Screening addresses quality. Usually, most of the quality is determined by the amount of manufacturing defects found in a lot. The testing, as it is stated above, intends to detect those defects. It is agreed that manufacturing defects may result in failures. However, there is more than one way to address the issue of the manufacturing defects. Another way to deal with the defects is by attempting to prevent them.
The present EEE COTS components are manufactured in a rigorous statistical process controlled high-volume production regime. That results in a substantial better outgoing components' quality. This questions the need for 100 percent screening.
The term "guaranteed" is inappropriate in context of the terms "level of quality" and "level of reliability". The term "guaranteed" is inappropriate in context of the terms "level of quality" and "level of reliability". To be on the safe side, it has to be mentioned that reliability cannot be tested into EEE components. It is understood that the testing as per this document is intended to increase the user confidence in the tested PEMs.
Quote from section 3.0: "There are numerous data indicating that improper handling and testing of the parts can introduce more defects than are screened out."
Quote from section 8.0: "Below are some additional guidelines for measures, which should be undertaken to avoid introduction of latent defects during testing, handling, and storing of the flight parts: Reduce handling by reducing the number of screening steps."
One of the more important decisions I made in the 90s, from the start of using EEE COTS Components, under a COTS first priority Selection Policy was: Do not test 100 percent the flight EEE components. Tested samples are not flown. A series of decisions resulted in a revolutionary methodology. The proof is in the pudding: multi-years, multi-satellites in LEO orbit, 100 percent success.
Quote from appendix A: "Screening of PEMs is essential before they are inserted into most flight hardware. The most important element in screening for reduced reliability risk for PEMs is burn-in."
In view of manufacturing COTS in a rigorous statistical process controlled high-volume production regime, resulting in a substantial better outgoing components' quality, in view of the potential damage to the test objects, in view of the costly infrastructure required for running and electrical verifying, it is worth to revisit the value of 100% burn in and the value of other post procurement 100% screening.
Quote from section 2.3: "For all PEMs, qualification by flight history or similarity is not acceptable."
It is not understood why NASA is still following the traditional way of thinking, imposing exclusively such a general restriction on all PEMs.
Automotive EEE COTS components
It seems there is a tendency to consider the use in space of automotive EEE COTS components. That is OK, but do not penalize the use in space of industrial level EEE COTS components. Do not exaggerate the value of the automotive added attributes in context with their relevancy to space applications. The goal is to select the "GOOD ENOUGH" and not "THE BEST". Thus should not be a competition of who is better.
AEC issued AEC - Q100, “Failure Mechanism based Stress Test Qualification for Integrated Circuits,” is the leading document for automotive ICs. There are different used terms to promote the sale of automotive EEE COTS components: "AEC-Q100 qualified", "AEC-Q100 certified", and "AEC-Q100 compliant". This may mislead the buyers.
Quote:
"Para. 1.3.1: "Successful completion and documentation of the test results from requirements outlined in this document allows the supplier to claim that the part is ‘AEC-Q100 qualified’." Para. 1.3.2: "Note that there are no ‘certifications’ for AEC-Q100 qualification and there is no certification board run by AEC to qualify parts.Each supplier performs their qualification to AEC standards, considers customer requirements and submits the data to the customer to verify compliance to Q100."
A NOTICE in the document states: "No claims to be in Conformance with this document shall be made unless all requirements stated in the document are met." Do not mix the terms.
AEC-Q100 requires an one-time qualification a new device family. No periodic qualification verification is required. The use of generic data to "simplify" the qualification process is strongly encouraged. Quote (Para. 3.1) "For each qualification, the supplier must have data available for all these tests, whether it is stress test results on the device to be qualified or acceptable generic data."
There is no qualifying activity to certify whether the EEE component manufacturer meets the requirements of AEC-Q100.
The EEE component manufacturer has the authority to self-approve his products as "AEC-Q100 qualified".
Each User is responsible to review the qualification data to verify compliance to AEC-Q100.
Screening is not required.
Automotive grade vs. industrial grade
The above highlighted elements of the AEC-Q100, points to a professional document containing a baseline of minimum requirements for qualification. Successful test results after performing all the specified one-time qualification addresses robustness/quality. However, the reliability is not addressed.
An AEC presentation states in context with disadvantages of the stress test qualification methodology (applicable to AEC-Q100): "cannot measure reliability, only can say the test passed (robustness)."
The AEC-Q100 contains a set of tests well known in the semiconductor domain. Consequently, most of these tests (if not all), at applicable test conditions, are applied also for any standard industrial grade EEE COTS component.
The most outstanding automotive tailored parameter is the extended operating temperature range. One of the main issues addressed in the AEC-Q100 is the tests adaptation to the relevant automotive operating temperature grade.
The narrower industrial operating temperature range -40°C - +85°C is in general enough for a space application (always there are exceptions).
The intimate interaction between the EEE component manufacturer and the automotive user, as mentioned in the AEC-Q100 is wishful thinking for a space user.
There is no argument that the controlled qualification baseline documentation in writing has his own value. Nevertheless, it would be worth to know and better understand the difference between an industrial-grade EEE component and an automotive-grade EEE component in terms of quality and reliability and not in terms of operating temperature range. Not having the prerogative of an automotive user (at least as stated in AEC-Q100) to access component manufacturers' relevant not published information, it is rather difficult (in diplomatic wording) to find explicit production flows for both grades.
A lot of argumentation is available to claim better quality and reliability for automotive EEE Components. The term "AEC-Q100 qualified" is the strongest, most impressive argument. If a "simple" user tries to assess what is done beyond the above one time qualification requirements, he will have to resort to general routine wording/slogans. If one wants to find out what are those advertised “special measures,” “extended measures,” “best commercial practice,” etc., he will probably end up frustrated (like me).
"AEC-Q100 qualified" catalog EEE components COTS meet the qualification requirements, but it may be that the automotive users are not purchasing them as is.
Many large volume automotive electronic system manufacturers DO NOT buy “catalog” automotive grade EEE components. Instead, they procure via internal SCDs based on “AEC-Q qualified” catalog items. Consequently car heritage is not an indication of reliable components.
When comparing datasheet of an automotive component versions vs. the industrial version, the operating temperature range is the main difference. However, the electrical parameters (test conditions, limits) may differ as well due to testing for wider operating temperature range.
In most of the cases both automotive and industrial versions of the same component are offered. It does not make sense for the EEE component manufacturer to design and manufacture two different die versions. The die dominates the component reliability. In other words, the automotive version is an uprated version of the industrial version or the industrial version is a downrated version of the automotive version. In most cases (if not all), both versions have identical dies manufactured on the same production lines. The built-in reliability, in such a case, is the same for the both versions. Claims at the workshop that special measures are taken for automotive components probably does not refer to die level. As mentioned above it is not clear (missing, not obtainable information) whether those special measures are value added for space application.
Obsolescence of EEE COTS components
A component becomes obsolete when the original component manufacturer stops marketing it. Electrical, electronic, and electromechanical (EEE) components obsolescence hits the military, space, and commercial market. In the military, massive EEE components obsolescence occurred due to manufacturers’ business decisions to leave or reduce activities in this market. It is interesting to find out that the resistance to change party members are sometimes frightening with obsolescence of EEE COTS components. They use for military/space level components the term "diminishing manufacturing sources" or "discontinued" rather than the term "obsolescence!"
It is out of scope for this article to tackle all the issues related to use of EEE COTS in space. Consequently I have to end it.
Let me give to a famous man the last word about the urgent need to make transition to the era of use of EEE COTS components, transition slowed down by resistance to change.
