Updates in silicon and electronics technology.

Ed.: This is a special feature courtesy of Binghamton University.

IMEC and Intel researchers develop spintronic logic device. Spintronics is a budding path in the quest for a future beyond CMOS. Devices use much less power than their CMOS counterparts and keep their data unpowered. IMEC and Intel researchers have created a spintronic logic device that can be fully controlled with electric current rather than magnetic fields. An electron’s spin generates a magnetic moment, and when many electrons with identical spins are close together, their magnetic moments can align and join forces to form a larger magnetic field. Such a region is called a magnetic domain, and the boundaries between domains are called domain walls. A material can consist of many such domains and domain walls, assembled like a magnetized mosaic. (IEEC file #12091, Semiconductor Digest, 1/21/21)

Plasmonics: A new way to link processors with light. Plasmonic transceivers transfer large amounts of data between processors. Fiberoptic links are the main method of slinging data between computers in data centers. Silicon photonics components are large in comparison to their electronic counterparts because optical wavelengths are much larger than transistors and copper interconnects. University of Toronto and Arm researchers have developed new silicon transceiver components that rely on plasmonics instead of photonics. The results have transceivers capable of at least double the bandwidth, while consuming 33% of the energy and 20% of the area, and could be built atop the processor. (IEEC file #12097, IEEE Spectrum, 1/21/21)

To continue reading, please log in or register using the link in the upper right corner of the page.

Signals and energy move in the spaces, not in the traces.

When a principle of physics is accepted, it explains phenomena everywhere in the universe. The law of gravity works on matter, whether the masses are located on earth, in the sun and or among the stars. The law must transition at the atomic level where the particles must follow the laws of quantum physics. There may also be a transition when dimensions are those of the entire universe. For the world, we can sense there is only one law.

The laws I want to talk about are the basic laws of electricity. I am not referring to circuit theory laws as described by Kirchhoff or Ohm, but to the laws governing the electric and magnetic fields. These fields are fundamental to all electrical activity, whether the phenomenon is lightning, ESD, radar, antennas, sunlight, power generation, analog or digital circuitry. These laws are often called Maxwell’s equations.

To continue reading, please log in or register using the link in the upper right corner of the page.

Commonly available tools can determine the relationship between gold thickness and plating time.

Regression analysis utilizes the relationship between quantitative variables to predict one variable from one or more other variables. These relationships are either functional or statistical. Functional relationships have perfect fits. For example, ice cream cones cost $1 each, so one can buy 10 ice cream cones for $10 or 20 ice cream cones for $20. Statistical relationships typically do not have perfect fits. For example, cholesterol levels and body weight have a non-perfect fit. It is commonly accepted that the term “regression” describes the statistical, not the functional, relationship between variables.¹

A regression model is a formal means of expressing the general tendency of a dependent variable (Y) to vary with the independent variable (X) systematically. The independent variable (X) is also referred to as the predictor variable. George Box (2007) states, “all models are approximations. Essentially, all models are wrong, but some are useful. However, the approximate nature of the model must always be borne in mind” (p. 414), implying there are random errors present, hence the nonperfect fit.^1,2 The expression of the general tendency of a dependent variable (Y) and an independent variable (X) is potent in many different fields.

To continue reading, please log in or register using the link in the upper right corner of the page.

Machine and operational costs have shrunk over the past decade.

For years, the word in the electronics industry has been laser depaneling is expensive. This may have been true for investments in laser machines a decade or more ago, but the situation looks different once operating expenses are accounted for, especially with newer systems. In fact, according to our data, depaneling with laser systems is the most efficient method for a range of applications, and the cutting results are excellent, which means quality standards are also met.

The trend in the price-to-performance ratio for current laser systems, especially with respect to production of rigid PCBs, is obvious: The cost of depaneling based on the effective cutting speed has fallen to approximately one-tenth of what it was a decade ago (FIGURE 1). This dramatic shift is based on three major factors, all based on the rapid advances in laser technology. First, capex cost for laser depaneling systems has decreased to almost 30% of what it used to be a decade ago. Second, overall throughput has improved more than seven times. Finally, the operational costs for energy and maintenance have noticeably decreased.

To continue reading, please log in or register using the link in the upper right corner of the page.