|TruView 180HDx X-ray inspection with embedded reel-to-reel inspection system. |
Radiography (x-ray inspection) is a technique that is central to all recent and upcoming counterfeit detection standards, including IDEA 1010B, CCAP 101, AS5553, AS6081, and AS6171. X-ray inspection provides the unique ability to "see" what is inside an electronic component without damaging it.
A top view x-ray of a real plastic molded component has dark regions in the x-ray image that represent dense areas in the component. Conversely, the light areas represent light areas in the field of view. For this reason the area around the component is represented in white. The x-rays traveling through the different density areas of the component under inspection cast a shadow onto the camera. Thus this x-ray imaging technique is also known as a shadowgram. A common technique used to identify counterfeit components using x-rays is to utilize an exemplar as a basis of comparison. The exemplar samples can be obtained in different ways. The usual method includes the comparison to components in previous lots that were obtained from trusted suppliers. In the event that such a priori information is not available, it is possible to compare the part to another one currently in use. In this case, it is often necessary to x-ray the printed circuit board with the assembled component. Both options, if available, must be used carefully, as manufacturers can change the leadframe structure, die size, and wire bonding schemes without notice. Therefore, it is imperative that the user seek more information on the part before judging it to be suspect.
The major challenge in determining a component's counterfeit status is the usual lack of an exemplar that can be used as a basis of comparison. The most common strategy used in this case is to perform the comparison with parts of the same lot. This in-lot comparison can be very powerful, because all parts within the lot must be identical. However, oftentimes counterfeiters re-mark different parts (that do not have identical x-ray images) to fulfill an order. Even if you do have an exemplar available, it is also common to find counterfeit components mixed with good parts. That is the counterfeiters' attempt to circumvent detection by customers who perform tests of just a few "sample" parts in the lot. For this reason, it is imperative that all parts are tested to assure homogeneity within a lot.
|Photo of a section of the reel, (left) all components showing identical part number, date code, and part number and (right) overlaid x-ray image onto photo showing mixed components. |
Until fairly recently, testing thousands of components was cost-prohibitive. However, recent breakthroughs in the automation of radiography inspection have made it possible and feasible. New systems in the market make possible reel-to-reel automated inspection of thousands of parts in trays, tubes, or reels.
When looking for counterfeits, the most common things you may find inside the components are:
- Inconsistent die size; because the die is a thin piece of silicon, a top view x-ray image will likely not show the die. However, the die attach fillet is dense enough to appear in a top view x-ray image. As a result, the size of the die can be determined by measuring the boundaries created by the die attach fillet. Since counterfeiters often re-mark different parts to pass as the part wanted on the market, it is almost inevitable they will mix in different components with different die sizes. Another technique to verify the size of the die is to take a side view x-ray image of the part. In this case the through density of the die is enough to appear in the x-ray image.
- Inconsistent leadframe. Similar to the previous case, counterfeiters often mix parts with different leadframes in the same lot.
- Broken or missing wire bonds. A broken wire bond may be a result of extreme mechanical or thermal stress applied to the part. Similarly, the absence of wire bonds is the sign of a suspect lot of parts. It is very important to note that some parts are packaged with aluminum wire bonds. Since aluminum is a low-density material, these wire bonds will not show in an x-ray image. For this reason, if you cannot see the wire bonds in a part, we recommend decapsulation it to confirm that they are not Al wire bonds. X-ray can show gold and copper wire bonds.
- Incorrect wire-bonding diagram. Even if the parts have the same leadframe and die size, you must verify that the wire-bonding diagram is consistent.
- Missing die. Although you may not be able to see the die directly, it is important to verify the presence of wire bonds and the die attach fillet. We found several examples of empty packages being sold as functioning parts.
- Inconsistent die attach voiding. The mature semiconductor manufacturing process leads to consistent parts. The presence of large variations in the die attach voiding may be a reason to fail the parts.
Same Part Number
In some cases, legitimate components with the same part number will have different internal structures. For them to be legitimate, the lot number and/or date code must be different.
The following three examples illustrate how good components can end up with different internal structures as seen by the x-ray inspection. In these examples we assume that the parts analyzed with the x-ray inspection have different lot numbers and/or date codes. Otherwise they should be deemed highly suspect.
- Different Leadframe Structure. Market globalization has led to the distributed manufacturing of electronic components. Therefore, it is common for original component manufacturers (OCM) to fabricate the same part number in their domestic and international facilities. It is entirely possible that these different locations procure different leadframes due to relationships with their suppliers. It is also possible for a manufacturing plant to change leadframes at a certain point in time based on design and/or supplier changes.
- Different Die Size. "Die shrink" is a term often used in the semiconductor industry to refer to the continuous miniaturization of integrated circuits. According to Moore's Law, the number of transistors in an integrated circuit doubles every two years. This means that the same circuit can be implemented in a fraction of the size of the previous generation of semiconductors. Thus, die shrink is a natural consequence of the evolution of the electronics industry. What it means to you is that it is possible to find legitimate components with different lot numbers, date codes, and die sizes. The wire bonding diagram of these parts is usually the same, but not necessary always identical.
- Different Wire Bonding Diagram. As a consequence of die shrink, the power and signal characteristics of different generations of the same component can change. For example, it is possible that the earlier generation of a part needed more power to bias larger transistors. Thus, the multiple wire bonds needed to carry the necessary current to bias an earlier version of the die may not be necessary for the lower power later versions of the same circuit. These examples illustrate that the determination of the authenticity of a component should not be done lightly. If the x-ray inspection shows that the parts have different die sizes and/or leadframe structures and the same date code and/or lot number, you can discard them as highly suspect. However, if the date codes and lot numbers are different, this does not necessarily mean that they are authentic. Instead, this situation requires more research to verify the authenticity of the parts under scrutiny.
The Inspection Outcome
Counterfeiters operate a lucrative criminal enterprise. The following two examples attempt to put numbers to assess the scale of this illegal industry. The reports shown describe the total amount of parts inspected within each lot, the total inspection time using a TruView 180-3HDx equipped with a Reel-to-Reel system. We also placed a monetary value to quantify the potential loss when selling or buying such parts. The cost per part included in this analysis reflects the prices we were quoted by different distributors at the time this paper was being written. Thus, this cost does not account for the potential loss of product if the parts were actually used in the production line.
Part inspected: Kemet Tantalum Low ESR Capacitors. The Total number of parts inspected: 2,149, with a total inspection time of 15 minutes. Cost the parts: $6.44 ea., and the reel's total potential loss, $13,840.
In another example, the Samsung K6X1008C2D CMOS SRAM was inspected. The total number of parts inspected: 3,000, with an inspection time of 19 minutes. These parts cost $31.52 ea. meaning a total potential loss of $94,560.
Contact: Creative Electron, Inc., 253 Pawnee St., San Marcos, CA 92078 760-752-1192 fax: 760-752-1196 Web: http://www.creativeelectron.com