Simple RFID tags evolve into complex ultra-low-power SoCs, NXP expects

The world increasingly turns to microelectronics for weapons to fight intrusion, fraud, and counterfeiting. Access protection for buildings, networks, and devices; authentication to block forgery of documents, components, or even food ingredients and pharmaceuticals; and more benignly, applications such as manufacturing workflow control, inventory management, and micro-commerce: all are seeking silicon solutions.

This growing need is bringing together two technology threads within the normally-quiet Austrian engineering team of NXP Semiconductors, according to Heinze Elzinga, director of product management for the NXP Identification Sector. One of these threads began in high-end markets such as banking, where NXP has provided authentication solutions based on powerful cryptographic processors. "Now, we see this authentication technology proliferating into more cost-sensitive markets, such as authentication of goods to prevent forgery," Elzinga says.

The other thread began in the thriving European enthusiasm for all things contactless, from RFID tags to contactless transit passes and passports to contactless smartcards. The need for remote sensing gets tangled up in this picture as soon as authentication technology gets applied to a fast-moving stream of people or a warehouse full of inventory that no one wants to scan one piece at a time.

Yet a third technology is just beginning to enter the picture: sensor electronics. In simple cases, an RFID/authentication chip may keep track of a single quantity: for example whether or not a product has exceeded a certain temperature or humidity during its lifetime. Or there can be more sophisticated measurements. For example, a tiny sensor network embedded in the packaging might monitor the pH of a piece of meat to detect spoilage. A more complex use case could involve biometrics, with, for example, an access card verifying that it was in its owner’s hand and not the equally-appreciative hands of the thief. This thinking can keep getting more elaborate, eventually reaching the concept of sensor fusion, in which a microcontroller in the chip assembles and correlates data from many sensors in an attempt to understand what is going on around it. Such an approach might, for example, detect an attack attempting to compromise the chip itself.

But all of these schemes must live within some very confining limitations. "These devices, especially in high-volume applications, must be monolithic for cost reasons," Elzinga says. That means that the RF communications, crypto processing, microcontroller, and sensor technologies all have to coexist in one process. And it means the entire design must be ultra-low-power.

"These chips can not be done with standard techniques," Elzinga maintains. "When this market says low power, it is talking in microWatts, not tens of milliWatts." Such designs must be rooted in energy-starving standards. They must employ custom—often hand-crafted—hardware, and very often still must rely on special process capabilities to meet the nearly zero-power budgets. "It helps to have your own fab," Elzinga suggests. "But even with control over the process, as the level of integration increases, the challenges seem to increase much faster."