Workshop Addresses Growing Dilemma of Fake Chips
By Chelsey Drysdale
Circuits Assembly
November 1, 2009

LOS ANGELES--The images are dismaying, but they are real. Workers in an electronics counterfeiting facility in China rinse chips in a murky river and throw them in piles on the sidewalk, separated by sizes. The smallest chips look like dirt mounds. These chips will later be marketed as the real deal.

Tom Sharpe of SMT Corp. displayed the photos he took on a recent trip to China to an incredulous crowd of some 300 attendees at the So Cal IMAPS/JPL Topical Workshop & Exhibition in Los Angeles in early October.

The staggering breadth--not to mention implications--of counterfeit components has led companies like American Electronic Resource to cease sourcing parts from China. AER has invoked a "No China" policy to defend itself against counterfeits. Before the company decided not to work with China, presenter Robb Hammond said, "50% of parts coming from China were counterfeit." That's right: half.

Presenters throughout the day discussed the growing problem of counterfeit products and offered steps to curb the issue. Rest assured, industry cannot rely on Customs to confiscate fake parts at the border. As Mark Goins of the US Customs and Border Protection department noted, customs' initiatives will only work if trademark owners are involved in border enforcement. "If they don't care, we don't care," he said.

That's not to say Customs hasn't taken notice. In the late 1990s, the agency became involved with Intel and AMD semiconductor anti-counterfeiting, recounted Goins. By 2003, the problem had escalated to a "much larger scale than had ever been seen." At that time, Customs approached trademark owners, who said that everything was fine. Consequently, customs' activity in this arena slowed in 2004 and 2005.

"2006 was the tipping point," he said. "Trademark owners couldn't ignore the problem anymore."

That year, the EU-US Summit on IP rights took place in Vienna to discuss joint border enforcement. The powers-that-be decided ICs were a health and safety issue.

As a result, at the end of 2007, a joint operation in five countries was launched for three weeks. In less than a month, there were 152 seizures, encompassing more than 400,000 devices. This involved 41 different trademarks in 15 ports in the US and EU, according to Goins.

"We decided we needed to continue doing this" beyond the three weeks, he said. "There were dozens of referrals to criminal investigators."

To date, nine people have been sentenced as a result of that operation, and many others await.

As of September, a 22-month operation has resulted in 999 seizures; more than 1.6 million semiconductors have been seized in 10 countries, involving 65 trademarks. And Goins believes they have only just begun.

Important to Customs' goals, he explained, are identifying promising cases for criminal investigation and providing trademark owners with information about all seized merchandise. "We are most concerned about multiple threats," he said, referring to health and safety issues as a top priority. "Big companies and big governments are all interested."

Finally, the issue is gaining traction at all levels and corners of the world. At the end of September, a World Semi Counsel meeting in Korea attracted customs experts, governments, and authorities from six countries to discuss semiconductors. "This year was the first year they talked about counterfeiting; the issue is being taken seriously," said Goins.

Customs may be taking the issue seriously, but two audience members directed pointed questions about border mistakes at Goins. They said their firms have experienced good parts being seized when red-flagged as possible counterfeits. Customs held those parts for months, the companies claimed, resulting in six-figure costs for the firm waiting for the parts' release. These are most likely isolated incidents, but worth attention.

What was clear from the workshop is that multiple organizations are trying to tackle the counterfeit problem. Some 10 years ago, a company had problems with a Toshiba part, only to determine later it was actually a Samsung component, which cost about a $1 less. The company reported its finding to ERAI, an electronics information services firm representing component makers and resellers. That was the firm's first experience with counterfeits, recalls Mark Snider, who works for the organization.

What's causing the counterfeit conundrum, according to Snider? The Chinese government overlooks the issue, he said, and there is an "enormous pool of cheap labor" there, which allows the labor-intensive activity of stripping e-waste and recycling the components. And the Internet plays a major role because it offers easy access to buyers around the world. According to Snider, "the supply chain is compromised, and needs a program to address the issue."

One step-by-step solution lies with an important standard. Art Ogg of World Micro suggested IDEA-STD-1010A, a protocol for detecting counterfeit components. A three-phase methodology, the standard aims at decreasing return material authorizations (RMA), while increasing returns-to-vendor (RTV). Implementing the standard should take about three months, he said, not including time spent to disqualify substandard suppliers.

IDEA-STD-1010A was written by independent distributors, and an update should be published by early 2010.

Both the standard and several companies support inspection. Yet Customs officers inspect only for trademarks on packaging and products. Says Mark Marshall of Integra Technologies, electrical testing will succeed where inspection cannot. "There are two classes of counterfeits: nonfunctional and functional," Marshall said.

Nonfunctional counterfeits are easy to discover, while functional ones are harder to detect. Inspection is only a first step, he believes.

Ticking off a variety of methods, including external and physical XRF; mark permanency; internal visual; basic DC test; minimum functional test at 25°C; full spec with extended temps, and test and quality, Marshall said, "IC manufacturing requires electrical testing; assured avoidance of counterfeits requires electrical testing."

The three levels of electrical testing involve basic DC validation, which is "most commonly provided." At this stage, pins are checked, and ESD damage is noted. It is the cheapest and quickest test, but involves no functional evaluation.

Next is the basic functional test; for example, some instructions are run through a microprocessor at a typical temperature of 25[degrees]C. In another example, only three chips out of 200 passed testing at 125[degrees]C, but they all passed at 25°C.

The third is functional test with parametric testing. "This is most similar to what the manufacturer would do," said Marshall. "The majority of data sheet parameters are tested" here. The downside is that the costs in time and money are significant, a month or more, and up to $20,000, according to Marshall.

How are costs determined? Marshall broke it down:

Costs also "vary based on the complexity of the device. Ten pieces will cost almost the same as 100 pieces because they have the same setup times."

"Reputable test labs have nothing to do with size," he added. "Some aren't doing what they say they are. If pricing is really low, be suspicious."

Finally, Marshall suggested testing "a part just like it is used. Match the tester to the device, and don't test every parameter on the data sheet."

To the moon. In March 2009, fakes entered NASA's supply chain, according to Tim Trainer of the Global Intellectual Property Strategy Center. In his presentation, Combating Counterfeits: Legal Enforcement Tools' Available, Trainer said NASA recently told Congress "costs are overrun because of fake parts." NASA personnel and aerospace companies have found stolen manufacturing markings on parts, he added.

He said 18 USC 2320 addresses trafficking in counterfeit goods, instituting stronger penalties for knowingly causing or attempting to cause bodily injury by trafficking in counterfeit goods. However, he is not completely convinced of the validity of other proposed legislation.

A proposed US Senate bill (S.1466) would increase disclosure of information to IP rights owners upon detention of goods. It would improve customs records, but proposes "self-certification of low-risk importers," a label about which Trainer rhetorically asked, "Who wouldn't identify themselves as low risk?" He also said S.1466 would increase IPR enforcement staffing, but it is unclear how that would occur.

A competing bill, S.1631, would establish a national IPR coordination center, which would be redundant, Trainer said. The bill discusses "availability of samples" and would require "at least one full-time person at each port specializing in IPR." His reaction to this last point was dubious.

Trainer noted that in 2006, US Customs issued $136 million in fines, but collected only half of 1 % of those fines. "Where is the deterrent?" he asked.

Firms should ask themselves these questions, according to Trainer: "Do you have a legal basis to go after this issue outside of the US? Are foreign governments willing to do something?" He suggests using US Embassy staff, understanding foreign enforcement systems, and possibly using other potential sources, such as Interpol and the World Customs Organization.

A proposed Anti-Counterfeiting Trade Agreement (ACTA) has a target deadline of the end of 2010.

"We need policy action," he stated. This was indeed the theme of the day.