The Evolution of the Counterfeit Electronic Parts Threat

We have detected counterfeit parts in most electronic technologies including caps, diodes, resistors, memory, etc.

Having a known good sample at time of test is simply another data-point for us; it is great to have to collect and catalog information but is not necessary. When authentication testing is performed at SMT Corp, it is treated in a similar fashion as a forensic investigation. Thorough research is done on each test lot including consulting with the OEM, referencing historical data books, using databases such as IHS and SiliconExpert, and previous test data from our internal database of 20+ years of test data. When all data is collected, it is reviewed by senior members of staff, all whom have been with SMT for at least 10 years. 

Yes, there are three notable databases; GIDEP, ERAI, and the Anti-Counterfeiting forum.

We perform authentication testing by treating each lot under test as an investigation. We acquire all data points and perform diligence by using many sources of data. Having a known good sample at time of test is simply another data-point; it is great to have to collect and catalog such information but is not necessary.

Yes. For example, you can look at the number of years’ experience the company has, evaluation of their processes, lab tours, industry accreditations, and recent GIDEP reports if you have access to GIDEP.  Also, you can ask for samples of test reports to evaluate the thoroughness and industry reputation is a big one. Also, be skeptical of abnormally low prices and lead times. Or you can work with SMT Corp.

Depending on the source of material (SMT procured vs customer supplied), we are obligated to report such information to GIDEP, ERAI, and the Anti-Counterfeiting forum. However, we cannot report if we are testing customer supplied material.

Because our counterfeit mitigation process begins at quote, SMT Corp rarely encounters counterfeits of any type during our procurement. However, we have encountered many clones during the past few years when testing customer supplied material and performing R&D. As far as percentage, I’d estimate at least 25%-35% were clones.

Currently, clones are made up of simple logic, memory type product (NOR, serial, etc) and microcontrollers (ARM, etc). We know for a fact that FPGA clones exist, but we have not encountered any.

I don’t believe counterfeit detection standards would be the appropriate method to upscreen COTS and automotive components to be used for LEO and GEO space applications. It is a very interesting question though. One could envision specifying a set of requirements for upscreening components to more of an industrial standard based on the LEO and GEO environment. We’d be happy to discuss this further as SMT also provides up-screening services based on MIL-STD-883 and MIL-STD-750 and other standards.  

Simply because AS6081 Rev A removed the specific test methods defined from the AS6081 standard. Now it refers to AS6171 test methods. We have a whitepaper explaining this in more detail: What’s the Difference between AS6171 & AS6081

This is a very good question and one that is often thought about. The potential for counterfeit EEE parts making their way into authorized distribution channels. There have been discussions over the years within DoD about zero trust policies for any components not coming from US controlled sources. With some of the examples I cited early on in the presentation about IP theft of designs and manufacturing blueprints it’s reasonable to be concerned about the potential to produce chips with the same processes as the authorized foundry. Also, one form of counterfeiting defined by AS6171 is authorized fabs producing unauthorized quantities and then selling them into the open market. It seems to me the fabless OCMs who outsource their fabs to Asia are at a higher risk for having unauthorized production quantities produced and sold into the open market, perhaps with less quality control. Then the question becomes can these make their way into authorized channels as well? 

We catalog based on part number and date/lot code. Occasionally we see differences in the data which are exactly what you mention; variability in good material between lots. We take this into consideration when performing authentication testing on a lot of material.

We occasionally get OCM support, however its mostly from low and middle tier OCMs. OCMs like Xilinx and Intel never provide support. Some manufacturers have supported us to the extent of providing documentation from the 90s for material they remarked for a specific purpose.

Very good question and, unfortunately, the answer is no. While we do occasionally get full support of some low and middle tier OCMs, OCMs such as Xilinx and Intel never provide support.