📜 ⬆️ ⬇️

Cosmic rays - the cause of random computer crashes?



If your computer suddenly freezes, issued a "blue screen of death" or could not copy the file - do not rush to blame the manufacturer of computer equipment or buggy memory. Perhaps the cause of the malfunction is cosmic radiation. Such events are called “ single-event upset (SEU)”.

A single event violation is a change in the state of the electronic component caused by a single particle of ionizing radiation (ion, photon, proton, neutron, etc.) that collides with a sensitive system node, such as a microprocessor, semiconductor memory, or a powerful transistor. The change of state occurs due to the appearance of a free charge, which appears as a result of ionization inside or near the sensitive node of the system or a logic element, such as a bit of memory. As a result, the device gives an error. This single error is called a “violation as a result of a single event,” SEU, or simply a soft error.

Occasional crashes due to cosmic radiation do occur occasionally even on the ground, and the probability of their occurrence in planes at altitude and in near-earth orbit is hundreds of times greater. The higher - the more likely, because there is more rarefied atmosphere and weaker protection against cosmic radiation.
')
The effects of SEU may be different. For example, in a digital photograph one pixel may fall out. Nothing wrong. It’s another thing if, due to the cosmic neutron, the computer system of the aircraft is buggy - and it has to go for a forced landing. This really happened once to the C-141B Starlifter military transport aircraft, which experienced an occasional crash during a flight over the Sea of ​​Japan with more than 100 passengers aboard. During the flight, the plane suddenly fell on the right wing. The crew managed to straighten the list and land the plane. A subsequent investigation revealed that the microchip in the automatic control system suddenly gave false readings with the wrong bit - probably due to a collision with a neutron.

According to statistics, at a high altitude, about 1600 cosmic particles per second pass through every square meter of surface. That is, about 600 cosmic particles per hour pass through each square centimeter. Based on such assumptions, random failures may not be such rare events as it seems to someone.

At an altitude of more than 9000 meters, the intensity of the neutron flux is 300 times higher than at sea level. The probability of a violation as a result of a single event also increases. Unfortunately, there is no real protection against cosmic rays, so it only remains to rely on luck.


On October 7, 2008, the Airbus A330-303 airliner from Qantas Airways was on its route from Perth (Australia) to Singapore. At an altitude of 11,300 meters, a failure occurred in one of the three reference inertial units, as a result of which incorrect data was sent to the computer control system. For this reason, the plane abruptly went down, throwing up passengers who were not wearing seat belts. 110 out of 303 passengers and 9 out of 12 crew members were injured. Among the passengers, 12 people were seriously injured, and another 39 people went to the hospital. Among all the possible causes of the failure of the inertial unit, only the SEU remained inactive, the rest were recognized as “unlikely” or “very unlikely”. However, the Australian Transport Safety Board considered “insufficient evidence to assess the likelihood” that it was SEU that caused the failure.

Although the probability of a single failure on Earth due to cosmic radiation is 300 times lower than at an altitude of 9000 meters, sometimes the most unexplained events that occur with computer equipment are attributed to this phenomenon. For example, in 2003, an electronic voting machine in the city of Schaarbeek (Belgium) added 4,096 votes to one of the candidates for the election. An investigation revealed that this failure was caused by a change in one bit in the device’s memory. The cause was called cosmic radiation. Characteristically, the error was discovered only because the candidate received more votes than was possible. Otherwise, the failure would have gone unnoticed.

“It’s really a big problem, but it’s largely unnoticeable to society,” says Bharat Bhuva, a member of the Radiation Effects Research Group and a professor of electrical engineering at Vanderbilt University ( USA). This research group was formed in 1987, including to study the effect of cosmic radiation on electronic systems. Initially, the group was engaged in military and space systems, but since 2001 it has expanded its area of ​​interest to consumer electronics.

Although there are quite vivid examples of equipment failures, SEUs remain an extremely rare phenomenon. But experts note that electronic chips are increasingly used in various household appliances. The density of transistors on chips increases, as well as their number. Because of this, the likelihood of meeting with the "space crash" is growing every year. Electrical manufacturers are exploring the problem. For example, in 2008, Fujitsu engineers climbed a Hawaiian volcano to measure cosmic radiation at an altitude of 4,200 meters. There it is about 16 times higher than at sea level.

To protect against cosmic radiation, consumer electronics manufacturers are trying to use less sensitive materials and error correction codes (error-correction codes). In more expensive devices, duplication systems can be used.

Engineers, system administrators and programmers now have a great excuse to explain strange glitches of computer equipment.

Source: https://habr.com/ru/post/401681/


All Articles