Fullscreen HTML5Med Res (??MB)HTML4

Radiation and electronics There are natural and artificial sources of radiation all around us, including nuclear radiation, the solar wind, cosmic rays and electromagnetic pulses. Radiation can have adverse effects on electronics, including critical electronics such as in aircraft, spacecraft and life support systems. It is therefore vital to understand the sources and consequences of radiation events in electronics, and how to prevent radiation from affecting electronics, or manage the results adequately, if that is not possible. by Dr David Maddison 14   14 14  S   S Silicon Chip Australia’s Australia’s electronics electronics magazine magazine siliconchip.com.au Fig.2: the Van Allen radiation belts comprise two or three regions of energetic charged particles (eg, electrons and protons), mostly from the Sun, which are trapped in Earth’s magnetic field. This diagram shows the location of the inner belt, the outer belt and the position of various satellites. There is a so-called “safe zone” between the inner and out belts which is relatively low in radiation. Image credit: NASA. R adiation effects on electronics are primarily of concern in aerospace and military applications, although not exclusively so. Ground-based computers also suffer from radiation-based errors regularly. This problem has been exacerbated by the continuous reduction in transistor sizes as higher computing speeds and lower manufacturing costs are required; smaller transistors are more susceptible to radiation effects. Space is one environment where environmental radiation is a major problem for electronics. The types of radiation encountered in space vary enormously with time and locality. Even if a spacecraft remains within a certain area, eg, the surface of the moon, low earth orbit or geosynchronous orbit, the radiation it is exposed to can vary greatly. This is influenced by factors such as solar radiation, which varies all the time, and “space weather” in general. It is not just the intensity or energy of radiation that changes but also the Opposite: an artist’s concept of the NASA Lunar IceCube spacecraft to be launched in 2020. It is a 6U CubeSat that uses a Space Micro Proton400K radiation-hardened single board computer (Fig.1; inset). Image credit: Recentcontributor2000. siliconchip.com.au types of radiation particles that are encountered. And these, in turn, affect both the likelihood and severity of effects on electronic components. Radiation can cause a variety of impacts to electronics devices, including long term degradation of devices like solar cells, loss or alteration of computer memory contents, halting (“crashing”) of computer systems (possibly requiring a reset) or causing computers to issue incorrect instructions. In severe cases, the entire electronics system or subsystem can burn out, rendering a system permanently inoperative. Electronics may be irradiated by particles such as electrons, protons, neutrons and ions as well as photonic radiation such as gamma rays and x-rays. Electromagnetic pulses (EMPs) can also cause problems. These can arise from nuclear explosions, lightning or other events which cause an electric or magnetic field or an induced electric current. Apart from the space environment, electronics may be subject to radiation in applications such as nuclear reactors (eg, control systems), particle accelerators, high-altitude aircraft, highaltitude balloons, x-ray machines, food irradiation machines (for preservation) Australia’s electronics magazine and radiotherapy machines for medical applications. Sources of radiation Some potential sources of radiation, the particles produced, and the effects they have are: • Cosmic rays – these are very fast particles which come from all directions in the universe. They consist of about 85% protons, 14% alpha particles (helium nuclei), 1% heavy ions as well as x-rays and gamma rays. Most of these are filtered by the atmosphere and therefore mostly spacecraft are affected; however, collisions between cosmic rays and particles in the Earth’s atmosphere can also generate secondary radiation which can reach the surface. • The Van Allen radiation belts surrounding the Earth contain electrons and protons, mostly from the Sun, which are trapped by the Earth’s magnetic field. The strength of the radiation in these belts varies enormously. Spacecraft are affected by them, and they are also hazardous to astronauts. (Fig.2) • Solar flares eject particles such as protons and heavy ions as well as x-rays, some of which reach the Earth’s atmosphere. These can be associated with solar storms or geomagnetic storms. July 2019  15 Fig.3: a proton or neutron impacting a semiconductor crystal lattice can displace an atom from its correct location and alter its electronic properties. Meanwhile, it continues through the crystal (with reduced energy), where it can potentially cause additional damage or electronic disruption. • Secondary particles can be generated by the interaction of primary particles when they enter electronic structures, eg, a cosmic ray which strikes the encapsulation of a device. • Gamma and neutron radiation is produced in nuclear reactors and can affect electronics inside a shielded area. • Particle accelerators such as the Large Hadron Collider produce various types of radiation that can affect unshielded sensors and control circuitry. • Nuclear explosions can produce a powerful electromagnetic pulse and a large variety of particles that can affect electronics and power grids. • Trace radioactive elements in electronic chip packaging and wafer materials were found to be a problem in the 1970s. Alpha particles (helium nuclei) in older packaging materials could discharge the capacitors in DRAM, but this effect has been minimised today by using purer packaging materials and more sophisticated error correction. Origins of damage or effects to electronic materials Radiation damage or effects to elec- Fig.4: a radiation particle, in this case an ion, passing through a field effect transistor (FET) structure. This can disrupt thousands of electrons. The flow of current passing through the structure is affected, possibly causing a malfunction in the circuit. The damage is usually temporary. Image courtesy Windows to the Universe. tronic materials may be either permanent or temporary while the source of such radiation can be in the form of neutrons, protons, alpha particles, ions, x-rays, parts of the UV spectrum and gamma rays. In terms of damage to electronics radiation can be divided into two main types. One type is high energy radiation which is capable of causing disruption of atoms in a device’s crystal lattice and permanent damage. The other type is that comprising of lower energy radiation that is not able to cause disruption in a crystal lattice but can cause disruption of electronic charge carriers in a crystal lattice. Permanent damage can be in the form of “lattice displacement” whereby atoms are moved from their correct positions, causing the formation of new electronic structures such as recombination centres, and worsening the properties of semiconductor junctions due to rearrangement of charge carriers within the crystal. Although such lattice displacement damage is usually permanent, in some cases limited self-repair is possible due to “annealing” whereby displaced atoms can move back or partially back to their correct locations. Individual instances of lattice displacement won’t necessarily cause noticeable degradation of a device. However, the effect is cumulative and multiple instances of lattice displacement cause long term degradation in the performance of a device. This could include, for example, alteration of the switching threshold voltage of a transistor, causing a transistor to remain permanently switched on or off, or reducing the output of a solar cell on a spacecraft. Another source of damage in semiconductor crystal materials is ionisation. The energy of particles involved in ionisation effects is generally too low to cause permanent damage but can create “soft errors” such as corruption of memory contents or alteration of circuit logic states (Fig.4). The damage can become permanent if a condition is generated such as a Single Event Latchup (SEL), which can lead to permanent damage under certain conditions (more on that later). Main types of radiation-induced effects Based on the above mechanisms, radiation effects in electronic structures can be broadly categorised as: Soviet ‘retro’ radiation hardening technology When a Soviet pilot flying a MiG-25 defected to the West in 1976, experts were surprised to find that a majority of its avionics were built with vacuum tubes. This represented old technology for the time, but it was concluded that the Soviet decision to use vacuum tubes was due to their better tolerance of temperature extremes than solid state electronics of the time. It was also considered that this meant that the avionics bays would not need environmental controls, and vacuum tubes were also more resistant to the electromag16 Silicon Chip netic pulse (EMP) from nuclear explosions than solid-state devices. Also, the tubes enabled the aircraft radar to operate at an extremely high power of 600kW. Having said that, at the time, the more modern electronics of the West was quite capable of withstanding adverse environmental conditions and EMP, so the real reason the Soviets used vacuum tubes was probably that their electronic industry was less advanced than that of the West. But there are still situations today where vacuum tubes are considered for use in space-based applications, because of their robustness. Australia’s electronics magazine siliconchip.com.au Figs.5 & 6: a Single Event Upset, whereby a heavy ion or proton passes through a memory element, creating electron and hole pairs due to ionisation within the crystal lattice. This creates a parasitic current which can alter the value of the bit stored in memory (a bit flip). In the case of a proton passing through the structure, secondary nuclear reactions can lead to further effects. Source: NASA. 1) Lattice displacement effects; described above 2) Total ionising dose effects; a cumulative effect of radiation causing long term damage 3) Transient effects, such as the short but intense pulse caused by a nuclear explosion which may or may not cause permanent damage 4) System-generated EMP effects which can result in destructively high currents 5) Single-event effects (SEE) – probably the most significant events electronics are subject to Single Event Effects SEE is the general term for a variety of phenomena such as the ionisation effects described above, in which a single energetic radiation event has some effect on the electronic state of an electronic structure. Single Event Effects can be classified as follows: Single Event Upset (SEU) – “soft” errors which result in no permanent electronic damage. SEU errors often manifest as ‘bit flips’ in memory, ie, a zero changing to a one or vice versa. In some cases, multiple bits can be affected. This can also result in inappropriate pulses in circuitry (see Figs. 5 & 6). SEU can potentially place the affected circuitry in some undesired mode such as a test mode, a program execution halt or some other unwanted state. An SEU can be cleared, if detected, by a computer or equipment reset, or by re-writing the affected bit with its original value, which was famously done in the Voyager spacecraft; see below. Single Event Latchup (SEL) – this can be either a “soft” or a “hard” error. A hard error can lead to the destrucsiliconchip.com.au tion of the device. In an SEL, a circuit element is forced into a high-current state, causing excessive heating beyond a device’s operational limits (see Fig.7). This could result in its destruction (hard error) unless the fault is quickly detected and the device is reset by power cycling. This type of effect was first noted in 1979, and it can be caused by heavy ions or protons. Note that the commercial radiationhardened chip (GR712RC) mentioned below has circuitry to monitor junction temperatures which can shut down and reset the device in this case. Single Event Burnout (SEB) – this is a “hard” error which destroys the device. Devices such as power metal oxide semiconductor field effect transistors (Mosfets) were thought to be the only ones affected by this, but it is now known that other devices such as power bipolar junction transistors (BJTs), insulated gate bipolar transis- tors (IGBT), thyristors, high-voltage diodes and CMOS PWM controllers and drivers are also susceptible. This destructive mode of failure is due to the passage of heavy ions or other particles, which may originate in solar radiation, through sensitive regions of the device. SEBs in power Mosfets have been known to occur in space-based electronics since 1986, but more recently, have been recognised as a possible source of failure for terrestrial devices as well. An SEB event occurs when a highvoltage semiconductor device is biased in an off state with a voltage close to its maximum rated value applied. A single ionising particle then strikes the depletion region of the device, generating a series of electronhole pairs. If the electric field in that region is strong enough, an avalanche or regenerative feedback effect is initiated, causing destructively high currents in the device. Fig.7: CMOS circuits contain parasitic bipolar structures which can be triggered by transient signals from radiation. Such circuits are protected by guard bands and clamps, but radiation signals can bypass these. Two parasitic transistors are shown in a four-layer device. If triggered, several hundred milliamps can flow, leading to rapid heating and destruction if this is not detected and stopped within milliseconds. SEL is more likely at higher temperatures. Figure courtesy NASA. Australia’s electronics magazine July 2019  17 Fig.8: the Fairchild Micrologic Type G three-input NOR gate from 1961, the first practical integrated circuit, as used in the Apollo guidance computer. During its manufacture, the price dropped from US$1000 to US$20, leading to its commercial use. It’s intrinsically radiation-resistant due to its large size and small component count (six transistors and eight resistors). To see how this chip worked and how it got humanity to the moon see: siliconchip.com.au/link/ aapx Only N-channel Mosfets seem to be affected by SEB; P-channel devices appear to be immune. Single Event Gate Rupture (SEGR) – this affects power Mosfets and is caused by the breakdown of the oxide layer on the Mosfet gate structure. The results are similar to an SEB event. Electrostatic charging of spacecraft Spacecraft can acquire an electrical charge due to their interaction with charged particles in space. Generally, spacecraft have a positive charge on the sunlit side due to the photoelectric effect, and a negative charge on the dark side due to plasma charging. This charge can occur either on the surface of or internal to the spacecraft. This can result in damage to electronic circuitry and interference with scientific measurements. Damage can occur due to electric discharges between adjacent components at very different potentials, or Fig.9: the RCA 1802, one of the first radiationhardened CPU chips. Image credit: CPU collection Konstantin Lanzet, CC BY-SA 3.0 siliconchip.com.au/link/aapy 18 Silicon Chip from an electric discharge due to an accumulated static charge within dielectric materials due to long-term bombardment with charged particles. The satellites most vulnerable to these effects are in geosynchronous orbit, where there is a low plasma density that does not allow a bleedoff of charge. Potentials as high as 20kV have been recorded. Spacecraft charging avoidance options are limited, but it can be mitigated by having charge dissipating surfaces, using design practices to minimise differential charging and careful consideration of spacecraft orbit and space weather during launch (eg, avoiding solar storms). Electromagnetic pulses Apart from nuclear explosions, electromagnetic pulses (EMP) can arise from lightning, electrostatic discharges, switching heavy current loads, non-nuclear electromagnetic pulse (NNEMP) weapons and electromagnetic forming, as used in industry to shape certain items. An EMP can induce strong currents in materials and damage or destroy them, wipe magnetic media, interfere with wireless communications, destroy national power grids and have many other adverse effects. Protection against EMP can include shielding and current limiting devices, but it is difficult to protect an entire power grid. Recognition of such a risk has lead to the US “Executive Order on Coordinating National Resilience to Electromagnetic Pulses” (see siliconchip. com.au/link/aapz). See also the report at: siliconchip. com.au/link/aaq0 It is not known if Australia has any specific plans to deal with such threats. Designing to minimise radiation-induced events Avoidance or minimisation of adverse events due to radiation can be achieved through appropriate component selection, digital error detection and correction, use of redundant components, detection of excessive current or heat at chip junctions (see Fig.11) and also shielding. The problem with shielding is that it is heavy and is also ineffective against cosmic rays. It can, however, be effecAustralia’s electronics magazine tive against solar flare particles. Components designed explicitly for radiation hardness are typically based on a commercial equivalent, with various modifications. They generally lag behind nonhardened devices in performance, partly because of the extra research, development and certification required to produce them and also because some radiation hardening features tend to lower performance. In fact, older, slower devices tend to tolerate radiation better due to their larger junctions, so ‘upgrading’ spacerated components is much more difficult than their terrestrial counterparts. In terms of susceptibility to radiation-induced effects, technologies in order of the least susceptible to the most susceptible are as follows: CMOS (silicon on sapphire), CMOS, standard bipolar, low-power schottky bipolar, nMOS DRAM (n-type metal oxide semiconductor dynamic random access memory). Radiation hardening of devices can be characterised as being based on physical methods or logical methods, such as error correction and redundancy. Physical hardening methods include: • fabricating chips on an insulating substrate such as sapphire, to reduce the possibility of parasitic stray current pathways caused by radiation events • the use of bipolar transistors in integrated circuits which use two types of charge carriers instead of FETs, which use just one • the use of SRAM (static random access memory), which is intrinsically more radiation-resistant than DRAM (dynamic random access memory), although it is larger and more expensive • the use of wide band-gap semiconductors such as gallium nitride and silicon carbide instead of silicon, which are less likely to be disrupted by a given electrical charge injection • shielding of electronics with materials such as aluminium and tungsten, despite the added weight • shielding of electronics with boron-11, which results in less secondary emission of radiation when struck by primary radiation Logical means of radiation hardening include: • the use of strong error correctsiliconchip.com.au Fig.10: the radiation-hardened Vorago RH-OBC-1 onboard computer and avionics board for spacecraft, specifically designed for CubeSats. • • • • ing schemes for memory, such as the BCH (Bose–Chaudhuri–Hocquenghem) cyclic error correction scheme. BCH (250, 32, 45) can provide 99.9956% correctness even with a 10% memory bit error rate (1 byte in every 711 would still be defective). Robust error correcting codes have a high computational overhead. the use of redundancy such as multiple redundant computers and software, as used on the Space Shuttle. With three or more computers, they can ‘vote’ if they do not all agree (see below) the use of multiple error correction schemes keeping multiple copies of critical information the use of a watchdog timer that will reset a computer if the expected behaviour does not occur after a certain amount of time Testing techniques Electronic components can be tested for radiation hardness by exposing them to radiation from sources such as particle accelerators, radioactive elements like californium and actual testing in space. The correct application of statistical techniques to determine true error rates is very important. Radiation and CubeSats CubeSats are popular, low-cost satellites often built on a tight budget and with commercial off-the-shelf (COTS) components. siliconchip.com.au Fig.11: the Ramon GR712RC, a radiation-hardened chip for space applications. It contains a dual-core LEON3FT SPARC V8 processor and was being used by the SpaceIL “Beresheet” lunar lander (see SILICON CHIP, November 2018; siliconchip. com.au/Article/11296). It uses Ramon’s proprietary “RadSafe” technology, with a dedicated design including circuitry to monitor radiation, monitoring of chip junction temperatures, error correction logic, hardened flip-flops, redundant circuit elements and a watchdog timer to reset of the chip if it crashes. The question is often asked if radiation hardening of CubeSats is necessary. The answer varies depending on the CubeSat mission, but in general, CubeSats have limited lifetimes in low earth orbit, where radiation is a much less serious threat than in other orbits. The limited expected life in orbit also limits the requirement for extensive radiation hardening measures. Radiation hardening in CubeSats is usually achieved through software, component redundancy and good component choices. A standard Android phone has been used as the control device on a CubeSat. On the other hand, the Lunar IceCube CubeSat mission to the moon uses a radiation-hardened computer – see photo on page 12. For more information on CubeSats, refer to the SILICON CHIP article on that topic in the January 2018 issue (siliconchip.com.au/Article/10930). Commercial radiation hardened devices, past and present As mentioned above, early electronic devices were less susceptible to radiation because of their large feature sizes. One such example is the Fairchild Micrologic Type G three -input NOR gate from 1961, as used on the Apollo guidance computer (see Fig.8). The RCA 1802 from 1976 (Fig.9) was one of the first microprocessors available in a radiation hardened version, fabricated using silicon on sapAustralia’s electronics magazine phire. It used the Complementary Symmetry Monolithic Array Computer (COSMAC) 8-bit architecture. The chip is still made today by Intersil, and sold as a high-reliability device, although its exact radiation resistance is unstated. It was and is used in the Galileo Probe, Hubble Space Telescope, Magellan spacecraft and various other satellites. The processor, in its bulk silicon version, was also popular with hobbyists. Further information on this chip is at the following links: siliconchip. com.au/link/aaq1 (device history) and siliconchip.com.au/link/aaq2 (regarding its use in amateur radio satellites). The Space Shuttle had a Data Processing System which comprised four IBM AP-101S General Purpose Computers with identical hardware and software, and a fifth computer with identical hardware but different software which had the same goals as the software in the other four computers. The computers would vote on any result, and any system in disagreement with the others would have its result excluded. While not explicitly stated, it is likely that this voting system took into account the possibility of data processing errors due to radiation events or for other reasons and the redundancy would ensure a correct result. A description of the system can be seen at: siliconchip.com.au/link/aaq3 Two current devices of interest that are radiation-hardened for space apJuly 2019  19 charged particles from the sun) then resulted in induced currents in telegraph wires, which caused shocks to operators and also started some fires. This storm was also known as the “Carrington Event”. The Aurora was seen as far north as Queensland. The original 1859 Moreton Bay Courier newspaper article about the aurora can be seen at: siliconchip.com. au/link/aaq4 2. The Starfish Prime Fig.12: a photo of the Starfish Prime nuclear explosion (400km altitude) taken 45-90 nuclear test: In 1962, the seconds after detonation in 1962. It caused an United States conducted unexpectedly strong electromagnetic pulse which high-altitude nuclear tests, destroyed several satellites and land-based detonating a 1.4 megatonne electrical devices. nuclear warhead 400km plication are the Vorago RH-OBC-1, above the Pacific Ocean, 1450km from designed for CubeSats (Fig.10), and Hawaii (see Fig.12). the Ramon GR712RC (Fig.11 The explosion caused an unexpectedly large electromagnetic pulse, reNotable radiation-induced sulting in electrical damage in Hawaii, events destroying 300 street lights, setting off Some notable events due to radia- burglar alarms and destroying a mition interacting with electrical ap- crowave link. paratus or electronics are as follows: Bright auroras were also observed 1. Geomagnetic storm, 1859: A geo- in the detonation area and in an area magnetic storm (also known as a solar on the opposite side of the Earth from storm) occurred on 1st & 2nd Septem- the detonation area. ber 1859. This resulted in numerous Apart from the electromagnetic sunspots and solar flares. pulse, the explosion also produced What is assumed to be today a cor- beta particles (electrons) which peronal mass ejection (the expulsion of sisted as an artificial radiation belt within the earth’s magnetic field until the early 1970s. The failure of many satellites was attributed to the energetic electrons injected into the Earth’s magnetic field by this detonation. These satellites included Ariel, TRAAC and Transit 4B, while the first commercial communications satellite (Telstar) was damaged, ultimately leading to its complete failure in 1963. The Russian Kosmos V satellite was also damaged, among others. A video about the Starfish prime explosion titled “Operation Dominic I and II - Starfish Prime Part 2 1962” can be seen at: siliconchip.com.au/ link/aaq5 3. Radioactive decay in electronics chip packaging: Errors from trace radioactive materials in electronics chip packaging and silicon came to be recognised as a significant problem in the 1970s. Alpha particles (helium nuclei) are a common result of radioactive decay but are sufficiently slow and massive that they generally cannot penetrate the housing of electronics (they are even stopped by clothing or a sheet of paper). However, alpha particles originating from that packaging itself can interface with and affect the electronics within. A very low alpha particle flux of 0.001 counts/hr/cm2 are required to minimise the problem. This is be- Finding out about “space weather” Spacecraft operators and operators of certain other sensitive equipment are concerned with anomalies caused to electronics by radiation. Radiation from space comes under the auspices of “space weather”, and several websites have been established where information on conditions can be obtained. Some such websites, including one from the Australian Government, are as follows: www.sws.bom.gov.au/Space_Weather www.spaceweather.com/ Videos on radiation hardening of electronics “Demonstration of the effects of radiation on a commercial video camera”: https://youwww.swpc.noaa.gov/products/seaesrt tu.be/5kE0Rsf9W_I * “Watch A GoPro Travel Through Extreme Fig.13 at right shows an example of space Radiation”: https://youtu.be/QZZR4DJLdfM weather data taken from the NOAA Spacecraft * “Declassified U.S. Nuclear Test Film Environmental Anomalies Expert System – #62”: https://youtu.be/KZoic9vg1fw (from 1962,Fig.13: a videospace about weather the effectsisofrelevant high alti-to spacecraft operation. This screen grab Real Time (SEAESRT). shows a space weather readout from the NOAA website, for a satellite in tude nuclear detonations) www.swpc.noaa.gov/ geostationary orbit at 270°E. 20 Silicon Chip Australia’s electronics magazine siliconchip.com.au Radiation-Hardened Atmel Range from As this issue was going to press, the following media release came across our desks. We’re not sure how many readers would be into space and satellite applications but we thought it interesting nevertheless! Designers of space applications need to reduce design cycles and costs while scaling development across missions with different radiation requirements. To support this trend, Microchip Technology Inc.has introduced the space industry’s first Armbased microcontrollers (MCUs) that combine the low-cost and large ecosystem benefits of Commercial Off-the-Shelf (COTS) technology with space-qualified versions that have scalable levels of radiation performance. Based on the automotive-qualified SAMV71, the SAMV71Q21RT radiation-tolerant and SAMRH71 radiation-hardened MCUs implement the widely deployed Arm Cortex-M7 System on Chip (SoC), enabling more integration, cost reduction and higher performance in space systems. The SAMV71Q21RT and SAMRH71 allow software developers to begin implementation with the SAMV71 COTS device before moving to a space-grade component, significantly reducing development time and cost. Both devices can use the SAMV71’s full software development toolchain, as they share the same ecosystem including software libraries, Board Support Package (BSP) and Operating System (OS) first level of tween 100 and 10,000 times less than the emissions from the sole of a typical shoe. 4. Voting error in Belgium: In 2003 in Schaerbeek, Belgium, there was electronic voting for an election, and a single candidate obtained an extra 4096 votes. The apparent error was only noticed because that was more votes than was possible. The error was blamed on a Single Event Upset (SEU) due to radiation, causing a bit flip (inversion of zero to one). To explain how this can happen, recall that binary code is represented as bits (zero or one) in positions for 1, 2, 4, 8 etc. Position 13 of a binary number represents a value of 4096. So if that bit flips from zero to one, for example, the binary number 0000000000000 (zero) will become 1000000000000 (decimal 4096). 5. Qantas QF72: On 7th October 2008, Qantas flight QF72 experienced two sudden, uncommanded pitchdown maneuvers at 37,000 feet altitude (11300m) which caused injuries siliconchip.com.au porting. Once preliminary developments are complete on the COTS device, all software development can be easily swapped out to a radiation-tolerant or radiation hardened version in a high-reliability plastic package or space-grade ceramic package. The SAMV71Q21RT radiation-tolerant MCU reuses the full COTS mask set and offers pinout compatibility, making the transition from COTS to qualified space parts immediate. While the SAMV71Q21RT’s radiation performance is ideal for NewSpace applications such as Low Earth Orbit (LEO) satellite constellations and robotics, the SAMRH71 offers the radiation performance suited for more critical sub-systems like gyroscopes and star tracker equipment. The SAMV71Q21RT radiation-tolerant device ensures an accumulated TID of 30Krad (Si) with latch up immunity and is nondestructive against heavy ions. Both devices are fully immune to Single-Event Latchup (SEL) up to 62 MeV.cm²/mg. The SAMRH71 radiation-hardened MCU is designed specifically for deep space applications. to passengers, crew and damage to the aircraft. Investigators traced the problem to one of three air data inertial reference units, which sent incorrect data to the flight control systems. The following causes were considered for the “upset” (as it is officially described): software corruption, software bug, hardware fault, physical environment, EMI from aircraft systems, EMI from other onboard sources, EMI from external sources and SEE (Single Event Effect). All were rated “unlikely” or “very unlikely” to have occurred, except for SEE due to radiation, which was rated as “insufficient evidence to estimate likelihood”. You can read the comprehensive and fascinating report about the upset at: siliconchip.com.au/link/aaq6 6. Voyager 2 bit flip: On 22nd April 2010, the spacecraft Voyager 2 (see SILICON CHIP, December 2018; siliconchip.com.au/ Article/11329) had a problem with the format of the scientific data being returned to Earth. On May 12th, engineers retrieved Australia’s electronics magazine a full memory dump from the Flight Data System computer, which formats the data to be returned to Earth. They found a single bit of memory had flipped to the opposite of what it was meant to be. They reproduced this in a computer on the ground and determined it gave the same data format problems as were being seen from the spacecraft. On May 19th, commands were sent to the spacecraft to reset the affected memory bit and on May 20th, engineering data received from the spacecraft was normal again. Interesting Videos . . . “Demonstration of the effects of radiation on a commercial video camera” siliconchip.com.au/link/aaq7 “Watch A GoPro Travel Through Extreme Radiation” siliconchip.com.au/ link/aaq8 “Declassified U.S. Nuclear Test Film #62” – from 1962, about the effects of high altitude nuclear detonations: siliconchip.com.au/link/aaq9 July 2019  21 How modern semiconductors are radiation hardened – by Duraid Madina Pretty much all modern processors are fabricated with a CMOS process, ie, with a chip made up of N-channel and P-channel Mosfets formed from doped semiconductor layers and insulating oxide layers, plus metal layers to form the wiring which distributes power and signals between the transistors. In CMOS devices, radiation can result in the accumulation of charge in the oxide layer, leading to a shift in the gate-source voltage for a given drain current. NMOS devices typically see a lowering in the threshold voltage, increasing current when the device is both off and on. PMOS devices tend to get ‘weaker’, ie, higher gate voltages are required to turn the device on, and when on, the drive strength is decreased. This is not the only way in which CMOS devices are degraded by exposure to high-energy particles: other processes tend to result in a linearisation of the drain current vs. gate voltage curve, which for both NMOS and PMOS devices leads to an increase in gate voltage required to turn the device fully on. These defects are effectively permanent and will continue until the transistor is entirely unusable. It is quite easy to measure this damage; techniques such as deliberately timing-critical ‘canary’ logic paths, structures such as ring oscillators, or even parameters such as the total power consumed by a device can be monitored during operation, with changes indicating impending failure. As CMOS circuits have continued to shrink in size, radiation strong enough to alter the electronic state of a circuit but not so strong as to permanently damage it has become a common concern. For a while, the decomposition of radioactive lead isotopes in solder joints was a significant source of single-event upsets, but these days, the dominant source of SEUs is exposure to cosmic radiation. The digital circuits most sensitive to single-event upsets are those for which a voltage is used to indicate the state by a multistable circuit, such as in the classic six-transistor SRAM cell, where a pair of coupled inverters store a single bit of information and are isolated when not in use. As the size of the four MOSFETs, the local interconnect, and the operating voltage has decreased over time, there has been a significant decrease in the amount of energy required for an energetic particle to change the state of such a bit cell. Non-array elements like latches and flip-flops, and other array memories including DRAMs and flash memories, are also susceptible. One way that the reliability of these cells has been increased in the face of radiation is to spread the transistor gates over wider areas to ensure that ion strikes affect only a single node potential rather than two or more. Fortunately, the decrease in size of CMOS circuits has also allowed an increase in complexity which can also be utilised to combat radiation-induced events. So in addition to lower level design techniques like the increased gate area mentioned above, it is also possible to add redundancy to critical flip-flop cells, or even add error detection and correction coding to critical registers. Higher level protection techniques can also be used, including active software- or microcode-driven ‘scrubbing’ of critical memory contents, replicating critical logic blocks to operate in lock-step, with majority vote comparators, or ‘stop and retry’ logic which causes the processor to recalculate any results where the veracity of the previous calculation may be in question. Where field programmable gate arrays (FPGAs) are used, or other chips with configurable logic blocks, it is also possible to perform ‘online’ reprogramming of any logic blocks where a fault has been detected. In chips where robustness is critical, designers even go so far as 22 Silicon Chip to add ‘fault injection’ logic. This allows the fault mitigation techniques described above to be more rapidly and thoroughly tested, compared to what is possible with typical lab-based radiation tests. An example: reliable instruction fetching One critical function in any microprocessor is instruction fetching. The processor needs a continual supply of instructions to tell each of the processor’s functional units what they should be doing at any point in time. It’s vital that this be done at high speed (otherwise the microprocessor might remain idle), but it is even more critical that this be done reliably, as a corrupt instruction could easily lead to a variety of different errors, including potentially subtle corruption of program state, rather than an immediate crash or hang. To meet the speed requirement, instruction fetching is typically performed with a hierarchy of logic blocks, each ‘closer to the action’ than the next. At the top level is typically a high-speed instruction cache, which stores a limited number of the most frequently executed instructions, eg, the bodies of frequently-called functions. If for any reason this top-level cache is unable to immediately provide an instruction to be executed, the result will be an undesirable stall of the microprocessor while the cache attempts to fetch instructions from slower cache levels, memory, or perhaps even a disk or network. Due to its limited size and speed-critical nature, radiation hardening of a top-level instruction cache frequently involves maintaining a completely separate copy. This copy is kept physically separated from the original to the maximum practical extent, to ensure that a radiation strike corrupts only one of the copies. For speed reasons, typically only the original is “plumbed through” to the processor’s core functional units, and an independent unit is tasked with checking that both the primary cache and its copy provide identical results. In case a mismatch is detected, a high speed “stop!” signal is asserted to pause the rest of the processor before a potentially incorrect instruction is executed. This remains asserted until a more complex mechanism (such as an error correcting code) provides a known-good instruction and restores this correct entry to both the original cache and the copy. This “stop!” signal is frequently one of, and sometimes the most speed-critical path in the entire processor. Given that it toggles relatively rarely, it is often implemented using special, power-hungry, high-speed circuit techniques. Moving away from the high-speed core of a processor, errorcorrection techniques which take correspondingly longer times to use are justified. As the size of caches and memories increases, making complete copies of these becomes less practical. So lower-level caches and main memories are frequently protected with modified Hamming codes where, for example, 64 bits of data are encoded into 72 bits so that the corruption of any two of the 72 bits can be detected, and the corruption of any one of the 72 bits can be seamlessly corrected. In a radiation-hardened environment, main memories are frequently guarded with additional, software-based scrubbers which continually calculate and recalculate checksums for instruction memory blocks, and compare those against known-good values. These blocks can be encoded with quite complex codes, needing thousands or millions of machine cycles to correct an error, but can be designed in such a way as to virtually assure recovery of the original data whilst still maintaining a relatively low overhead in terms of space required to store the encoded data. SC Australia’s electronics magazine siliconchip.com.au