Scripting a Counterfeit IC Detection Strategy
Imagine a squadron of ten fighter jets taking off from a carrier in response to a threat. As they climb higher, the temperature plunges into the high negative degrees Celsius range. On four of the jets their electronics fail due to counterfeit integrated circuits that were not rated for those extreme temperatures. This was a concern of the US Navy that led to a 2010 report on counterfeit electronics.
The report titled “Defense Industrial Base Assessment: Counterfeit Electronics” looked into the infiltration of counterfeit/defective electronics into U.S. defense and industrial supply chains. The report found that 39 percent of companies participating in the survey encountered counterfeit electronics during the four-year long study. Worse yet, the trend of counterfeit incidents rose from 3,868 incidents in 2005 to 9,356 incidents in 2008. These counterfeit incidents included multiple qualified components resulting in a potentially much higher count of fake parts. In dollars of lost sales, this would equate to as much as $1.2 trillion in 2009.
Many people may jump to the conclusion that solving the problem is quite simple, do not purchase counterfeit parts from the black market. But this report found that counterfeit parts were purchased from the original component manufacturers, distributors, circuit board assemblers, and government contractors. Because these entities buy back unused parts from companies, counterfeit parts snuck into the supply chain. And because government projects, such as a fighter jet, most often exist for many years past the manufacturing life of a part, obsolete parts are heavily sought after.
Some highlights from the report:
- “all elements of the supply chain have been directly impacted by counterfeit electronics”;
- “companies and organizations assume that others in the supply chain are testing parts”; and
- “stricter testing protocols and quality control practices for inventories are required.”
Fake IC Parts
The rising occurrence of counterfeit parts is leading to faulty products, increased cost of returns, and additional costs to monitor and test for faulty components. Counterfeit parts include discrete as well as microcircuits. The five most commonly counterfeited semiconductor types are analog integrated circuits, microprocessors, memory ICs, programmable logic devices and transistors, based on data from IHS.
A counterfeit may be a non-working unit, a used unit that is repackaged and sold as a higher performing unit, or a marginal unit. The incorrectly marked and non-working units are the easiest to detect. For example, a fake part may use a trademark logo instead of the correct registered trademark logo.
Today suppliers are taking additional steps to detect counterfeit parts before they are resold. The testing includes: visual inspection, x-ray, optical, acoustic, de-capsulation, and electrical testing. These techniques may find non-working parts, but they are not enough to detect a working part that has been relabeled and sold as a higher performing part. Uncovering this type of counterfeit part is much more difficult. In order to detect parts that do not meet the required specifications, such as being able to operate from -25°C to +25°C, additional testing solutions are required. The majority of original component manufacturers, distributors, and circuit board assemblers visually inspect parts but less than half electronically evaluate the parts. Electronic and performance testing is more costly.
Stricter Counterfeit Testing
Proposed is an automated solution for testing parts post-assembly, in addition to the screening that was performed earlier in the supply chain. Assembly is the last step in the process and allows one to verify that a microprocessor, memory IC, or programmable logic device both matches the design specification and has the ability to perform correctly through various environmental conditions.
The first step is to create a script that extracts data from the microprocessor, memory IC, or programmable logic device and validates that data against known good values. The script would pull part number, serial number, revision, manufacturer’s date, and register values from the part after the unit is assembled and powered on. The script would then compare these values to know good values and report if any discrepancies were uncovered. Using Kozio’s VTOS SCAN product, that script is created automatically by using a single “golden” circuit board, can be executed on every assembled unit, and execution time would vary but most likely take under a second.
The second step is to use an automated performance test sequence that can be executed on the unit while the unit is placed in an environmental chamber. While in the chamber, the performance test sequence would be executed for a given amount of time at a given temperature, and then repeated over the required temperature range. Using Kozio’s VTOS environment and test sequencer, this step is accomplished by using the built in functional tests as well as the scripting language for flow control. This particular step has been used by many Kozio customers to verify part performance through a specified temperature range.
These steps can be incorporated into an existing process for counterfeit part detection, and quickly adapted and reused on future projects.
On December 31, 2011, US President Barrack Obama signed the fiscal year 2012 US National Defense Authorization Act, which creates regulations for counterfeit part detection and avoidance. Members at all tiers of the defense supply chain must put counterfeit risk mitigation procedures in place. This write up discusses two steps that can be implemented by circuit board assemblers to provide additional counterfeit testing on active parts.
Later this year, please anticipate a white paper providing additional information. I’d love to hear what you think and continue the discussion.