September 30, 2011

VisionTech's Dirty Dealings in ICs

by Amelia Dalton

In this week's Fish Fry, I dig into the dirty dealings of component house VisionTech. I investigate how one company from Clearwater Florida managed to dupe over a thousand unsuspecting customers out of $16 Million dollars and how these counterfeit ICs have affected our industry.

I also have an Amazon.com gift certificate to give out this week, but you'll have to listen to find out how you can win.

Channels

Embedded. FPGA. MilAero. Products. Semiconductor.

 

Watch Previous Fish Frys

Fish Fry Links - September 30, 2011

More information about VisionTech (1)

More information about VisionTech (2)

More information about VisionTech (3)

Fish Fry Executive Interviews

Moshe Gavrielov, CEO - Xilinx

John Bruggeman, Former CMO - Cadence Design Systems

Darrin Billerbeck, CEO - Lattice Semiconductor

Lauro Rizzatti, Vice President of Marketing, EVE

Bill Neifert, CTO - Carbon Design Systems

Sean Dart, CEO - Forte Design Systems

Kapil Shankar, CEO - SiliconBlue

Andy Pease, CEO - QuickLogic

Rajeev Madhavan, CEO - Magma 

Comments:

amelia

How could one small component house in Florida fool over a thousand companies and net almost 16 million dollars in the process? Check out this week's Fish Fry to hear the whole story. Do you think we can stem the title of counterfeit chips? Posted on September 30, 2011 at 1:38 PM

dougwithau

Very interesting. The question is always how close is the copy? Are they the same part that failed inspection, non-mil spec relabeled as mil-spec. The close but, not exact copy is really scary. The other scary thing is how many parts are on the shelf, sitting in a bin waiting to be put on boards. Plus, as a small company we use contract manufactures who buy the parts for us. We don't really know if they got the wrong ones. I have heard horror stories of money spent on debug and testing, paid by the customer, not the contract manufacturer. I don't know if I have ever encountered fakes. I have experienced lot problems. RAMs that failed only sometimes, and only from one lot. A current source that was waaayyyyy off in a new batch of boards. Fake, or just bad luck? As engineers we tend to be honest and never imagine outright fraud. We figure it was a bad lot, and move on. It all comes down to thorough testing from design through manufacturing. That drives up costs for everyone. Plus, it takes time, which I don't have. What makes me sad about this is that it hurts the little guy. A small company can not handle the costs of complete in circuit testing and the design work needed to do it right. It puts up yet another barrier to innovation, because the innovation requires trust. Posted on October 02, 2011 at 12:05 PM

Lord Loh.

People should be able to feed in the part number, date stamp and fab location on the manufacturer's website to verify if the chip is authentic. While this is not a 100% fool proof, it can weed out a lot of fake chips. Serializing packaging units (tape, tube reel) to be verified from the manufacturer's site should be useful for high volume customers. Posted on October 02, 2011 at 9:57 PM

krotar

A couple of years ago I was responsible for building a batch of low volume boards that had been designed 6+ years prior. These boards had 4 CPLDs that had been obsoleted with no drop-in replacement. I went to the internet to see if I could find some stock somewhere - I found a distributor that specialized in obsolete / hard-to-find components that told me they could get my CPLDs. The components came and I shipped everything off for manufacturing. 60% of the boards came back dead. I traced the problem back to the CPLDs - the JTAG chains were broken and the power planes were shorted internally. The failing CPLDs all had the same marking on top - I checked with the manufacturers website and discovered that the date code listed was invalid. I wired up a test socket to verify that the remaining stock was not counterfeit. It turned out that the distributor had sourced my order from 2 different sources and only 1 of these sources was fake so I was still able to get my boards built. Going forward I would have to be pretty desperate to stray from one of the mainstream electronics distributors again. Mine is a pretty minor example - far from mission critical - but it really brought home for me that electronics counterfeiting is something none of us can ignore. Posted on October 03, 2011 at 7:02 PM

kmathis71

I've not knowingly faced these issues, but as others have stated, it is possible those "bad lots" were fakes instead. I see two related but different issues. 1) The supplier that has been contracted to make a particular volume of parts, but overproduces to sell some themselves. This type of fraud drives down the value of the contracted chips, but probably doesn't garner significant risk of getting terribly bad chips. 2) Then there are the groups that have nothing at all to do with a legitimate product, solely seeking to fake parts. Perhaps they give a stab at reverse engineering, but are they really trying all that hard? The less effort they put into it, the greater their profit. By hiding in nations where we have little legal recourse, they are largely immune to the consequences. There are a few steps we can take to reduce these problems. I've seen IP to provide "serialization" from the silicon itself, taking the particular path properties to generate a unique code. Such things would make the blatant fakes less viable, but ONLY if the industry takes to more thorough validation testing. Stopping the first issue is far more difficult. In this case, the silicon is a match with the real parts. While the manufacturing is done in countries where we have little legal rights to enforce contracts, we will be stuck with this issue, in my opinion. I think these issues are only likely to increase with time and the continued global demand for electronics. Demand will find a supply. We will all be affected as the costs to verify the chips drives up component prices and thus product prices. Posted on October 04, 2011 at 11:29 AM

kevin

I think a couple things will make this better over time: 1) consolidation in fabs - as we move to more and more advanced processes - production will be done in fewer and fewer fabs and things like overbuilding would be much more visible. 2) Increased integration into large SoCs/FPGAs, etc. - if your whole design is mostly in one chip - there are fewer other parts to fake. Countermeasures can be built into the one big chip more effectively. Testing and checking become simpler. Posted on October 04, 2011 at 3:44 PM

daveDNE

Seems like you always get in trouble when the decision is to try to stay with an obsolete part. Had an experience on a medical cataract surgery unit with a very obsolete microprocessor. I was leading a re-design effort and I kept advising that there was no decent supply chain for this microprocessor and we were sourcing these microprocessors from 'brokers' and other dubious suppliers. There were rumors of suppliers filling orders by removing parts from old boards. The cost implications of re-porting the software over-ruled any supply chain problems. When the design was moved to manufacturing the supply people were shocked to find how the design team was buying these components. (the product was manufactured). So, I don't think all potential problems are considered when the decision is to stay with obsolete components. My experience has been you are setting yourself up to be more vulnerable to shaky suppliers including counterfeit chips. Posted on October 05, 2011 at 12:35 AM

jburger

Amelia, you missed a key part of the story. If you have the US Attorney's sentencing memo see footnote 11, and read your collogue Bruce Rayner’s story discussing how US Customs’ redaction policy may have thwarted discovery of thousands of counterfeit VisionTech shipments alone. You asked how we stop the flood of counterfeits into this country – the vast majority of which come from eWaste “recycled” mainly in China, seehttp://www.businessweek.com/go/tv/counterfeit.... There is no one silver bullet, but there are at least five steps that would go a long way to ameliorating this serious problem. First, end the Customs redaction policy (this is a long story – let me know if you want more information –essentially allow Customs’ Officers to resume sending unredacted pictures of suspect counterfeit chips to the trademark owners who are the only entities in a position to identify counterfeits). Second, continue prosecuting rogue brokers like VisionTech found through the Customs seizure process. Third, the government and companies assembling critical devices must require their contractors and their subcontractors only to purchase critical components from authorized brokers. Fourth, responsibly recycle eWaste (we need the rare earths anyway). Finally, we have to put more pressure on countries – primarily China – to crack down on these backyard “recyclers” that are supplying the dangerous counterfeits. The sample list of some of products where companies purchased counterfeits from VisionTech will curl your hair – just one, controllers for high-speed trains is frightening enough – let alone numerous critical military products. It is bad enough we have to send our military men and women in harm’s way. It is inexcusable to send them with equipment that may fail in combat because of a counterfeit chip. Posted on November 01, 2011 at 9:02 AM