Depending upon your relationship with the supplier, e.g., you are a large volume,
consistent purchaser or influential member of the supply chain), you may have the right to approve or reject proposed changes to the supplier’s product. If not, and changes are important to you or your customer, you may have the option of requesting notification of changes. Keep in mind that this notification may be concurrent with the change or even after the fact.
Often the level of insight and data to which you will have access must be established up front in the contract or purchase order’s terms and conditions. If you specifically want automotive-grade parts, you must purchase them (likely through distributors, unless you are a large-volume purchaser) by requesting parts that meet the AEC standards. Otherwise, distributors will fulfill your order from available stock, whatever that might be, to meet your quantity, cost and schedule requirements.
If the success of your designs depends on consistency of parts, materials, processes or test protocols, then change notification is important. In addition, there may be some changes more important to you (e.g., wafer diameter, bond wire material or molding compound composition) than others (e.g., plant location, product marking). If your business relationship with the supplier permits it, determine that minimum level of change important to you and seek a notification agreement from your supplier. In some cases, your customer may require such notifications or have the right to weigh in on them.
Consumer’s Risk is the risk that a Producer’s screens were inadequate and failed to remove some percentage of defective parts. These parts get delivered to the consumer or marketplace.
Producer’s Risk is the risk that a Producer’s screens were too stringent and scrapped good product that could otherwise have been sold to a customer.
With alternate-grade parts that undergo no burn-in or screening, the Consumer’s risk is higher (i.e., the customer is more likely to receive some small percentage of defective product). Typically, COTS and industrial-grade parts undergo limited to no burn-in or screening. If the results of burn-in, screening or other test data are available, they should be reviewed so there are no surprises (anomalies, unexpected failures) during your manufacturing and test programs or during flight.
This practice is known as qualification by similarity. If the new part being considered is, for example, manufactured at the same facility by the same staff, using the same materials, processes and test approaches as the qualified part, it is likely to be acceptable for your mission under a similar flight environment. That said, if the part qualified by similarity is a single-point failure, in a single-string application or in a mission-critical application, additional mitigations may be needed to further reduce risk.
Advertised FIT rates should meet or exceed those established for the mission or
early failures could terminate the mission prematurely. As a mitigation, incorporating redundancy into the design may be an option.
It is essential that the reliability of new technology or new devices was established
with criteria that mimics or envelopes your own. For example, your criteria may require that the reliability testing occur with the parts powered on and zero failures allowed, while the supplier may test the parts in a powered-off state with a limited number of failures allowed. Otherwise, you may expose the parts to new environments for the first time during flight, which could impact performance or endanger the mission.
Understanding your parts’ failure mechanisms is crucial to the success of your
designs. Without this knowledge, you may inadvertently apply the part in a way for which it was not intended and incur the risk of anomalies or failure. This is especially important for new technology or for existing technology utilized in a new or different application for the first time. Search databases and academic literature for failure mechanism data and discussions, e.g., https://radcentral.jpl.nasa.gov/ or https://radhome.gsfc.nasa.gov/radhome/RadDataBase/RadDataBase.html
A few tests designed to demonstrate failure mechanisms are as follows. HTOL, high temperature operating life, is generally intended to surface diffusion mechanisms, metal migration and annealing processes. Electrical bias promotes several temperature-sensitive mechanisms that depend on electric fields or current. THB, temperature-humidity bias tests aim to accelerate galvanic and electrolytic corrosion and other chemical reactions, material delamination and charge separation of insulating surfaces. Temperature cycling thermally induces mechanical stress to precipitate failure due to material fatigue and workmanship. Vibration testing promotes fatigue failures.
Auditing a supplier can provide valuable insight into their manufacturing discipline,
process control, and quality control and assurance. Auditing can build confidence in the supplier’s capabilities and likelihood that they are delivering product that meets your mission’s requirements. It also give insight into how effectively the supplier can respond and resolve issues, should you receive defective product.
That said, audits take time and money to conduct, whether the supplier is on- or off-shore. You also need to know what to look for and ask for, if the audit is be worthwhile. Additionally, contracts, non-disclosure agreements and other arrangements may need to be established in order for an audit to occur.
Parts tested over a temperature range narrower than that of your application may
fail in in your mission’s environment. Similarly, parts tested over a wider range may provide performance margin or enable you to choose a less expensive or lower-grade part if the margin is not justified.
Creating your designs with performance margin or redundancy are two ways to
mitigate the risk that undisclosed changes to parts produced at different times will not have a negative impact.
Unless you procure parts to an SCD (specification control drawing), specify compliance with AEC standards, pay for a PPAP (production part approval process), acquire independent lab test service, test data yourself, or have a relationship with the supplier that includes such disclosure, there is no guarantee that parts procured today are identical to those procured in the future.
To a great extent, the specific grade of the device determines the amount of data and level of insight you will have. This data and insight helps users determine lot homogeneity, i.e., consistency within a given lot, and consistency between different production lots. Generally speaking, the insight and level of data provided is higher for automotive and military grades than for COTS and industrial grades. Designs should be able to accommodate variations in part performance. This is often accomplished by including margin in the design requirements. And access to lot screening data, for example, increases confidence that infant mortals have been removed from the population from which your parts come.
If the parts have survived the space environment in a previous mission, this
provides useful information that could build confidence in the parts for your own application. Remember that there may be specifics of your mission type that are more (or less) stressing that could require additional or different testing or analysis to assess performance and reliability. Also, compare the respective length of the missions, which may impact reliability requirements.
– Select parts that have radiation test data available from either the manufacturer or third party independent test facilities. There are databases (see sites above) of COTS type parts that have been tested for radiation tolerance. Many of these parts exhibit tolerance to the radiation environment and could be used.
– Select parts that have known space flight history (heritage). There are many parts that have been flown on other missions that may or may not have radiation test data available. These parts should be considered for use based on their flight heritage.
– Select parts that belong to radiation tolerant families – CMOS components have been shown to have typical dose tolerances greater than 10krad. These parts should be considered when there is no other part available with a known history.
– Spot Shield: Spot shielding specific parts can greatly increase total dose tolerance for parts that may show weak performance or do not have any test data available.