Customer Reviews

Single Event Phenomena

5 star:		(1)
4 star:		(0)
3 star:		(0)
2 star:		(0)
1 star:		(0)

 

 

In the past couple of decades, we've seen a dramatic increase in the density of active elements on logic and memory chips. The volume and charge associated with a memory or logic bit has thus been reduced. But there's a problem: at some point, a single cosmic ray can act as a signal to flip a bit. And that point has long since been reached for heavy-ion cosmic rays. Luckily, devices on the surface of the Earth are shielded by the atmosphere from most of these, except for the less-dangerous protons. However, that still puts space-borne chips at risk, as well as chips on Earth that are susceptible to protons. For the past 25 years or more, these bit flips have been called "single-event upsets" or simply "SEUs." This book shows how to analyze the problems these cosmic rays present. It's an excellent textbook, with good sets of problems at the end of each chapter (and answers at the end of the book).

The book starts with some important preliminary material on flux, fluence, current density, and cross-sections, as well as some valuable basics on chord distribution functions. Those chord distributions come in handy when you want to know how much energy particles will deposit in the SEU-sensitive region. Next is a chapter about the enemy: the nature, distribution, and flux of extraterrestrial SEU-inducing particles. After that, there's a discussion of the amount of energy deposition per unit track length, which leads to an explanation of linear energy transfer (LET). That leads to the introduction of a fundamental concept in SEU error-rate calculations, the Heinrich curve, which shows the distribution of cosmic rays by LET. It also warns the reader about placing too much reliance on this curve: SEU cross-sections can vary with ion type, high-energy ions are more dangerous than low-energy ones, even if both have the same LET, the energy lost by a particle may not all go into the nearby SEU-sensitive region, there are variances in the amount of energy deposited, additional energy can be pulled into the sensitive region by "funneling," and there may need to be corrections for cosmic-ray showering.

After that comes a chapter on how to test devices for SEU susceptibility, using accelerators such as Brookhaven or simply exposing them to Californium-252. These tests not only look at bit flips, but also at "latchup," which can cause permanent damage to a device if it is left powered on.

Next there is a detailed discussion of SEU rates (and techniques for computing them) for a wide variety of particle environments: geosynchronous altitudes, the Van Allen belt, neutrons at aircraft cruising altitudes, the South Atlantic anomaly, and sea level.

Well, what is to be done about all these problems? The authors finish by discussing shielding, circuitry-hardening, redundancy, scrubbing, error detection, and error correction to reduce or counter bit errors. That requires a discussion of multiple-bit upsets as well.

This is a very useful book for anyone who has to deal with such problems, especially in the design of space-borne systems.