The high cost of satellite design launches necessitates the need to design low-risk, space-qualified systems before exposing said systems extreme environment of space.
There are a number of considerations mitigate the risk of system failures in the early stages of design especially for communication systems.
Vibration during Launch
All assemblies and subsystems are subject to intense vibration during the launch of the spacecraft. Therefore, thorough vibrational testing is required of a majority of the systems and subsystems in Phases II and III of the program.
Temperature Variation for GEO
In the shade, the temperature can fall below -150C, but in the sunlight, the exterior temperature can rise over +150 C. This huge temperature differential highlights the design challenges that have to be dealt with.
The satellites are designed to manage the heat from the sun with multilayer insulative coatings and radiators to distribute the heat across the satellite. However additional care must be taken for the electronics and communication systems as they are not typically designed to withstand such temperature variations.
The most sensitive of electronics are encased in warm boxes to keep a constant temperature, and bandgap voltage references are designed with PTAT + CTAT behavior for constant voltage output to circuitry regardless of the temperature.
Radiation Sources
Outside the protection of Earth's atmosphere, satellites and their electronics are subject to a gamut of radiation sources: gamma rays, X-rays, electrons, protons, neutrons, and ions. The radiation environment is both static and dynamic with both solar and Van Allen radiation contributions amongst other contributors from the universe.
Figure 1: Earth's Magnetosphere and Radiation Belts
Figure 2: Earth's Van Allen Belts over orbital altitudes
Contributions from the Universe: galactic cosmic rays are high energy particles originating from anywhere in the universe
Solar Storms: solar flares and coronal mass ejections
Van Allen Belts: these belts are held in place by the Earth's magnetic field and trap charged particles carried by the solar winds.
A Geostationary Earth Orbit brings the satellite out of the range of heavier trapped protons residing in the inner Van Allen belts. However, they are still exposed to lower mass trapped electrons in the outer Van Allen belts. They are fully exposed to protons and heavy ions from solar flares, as well as galactic cosmic rays. During solar events, a GEO satellites proton exposure increases from coronal ejections.
Radiation Effects
The effects of radiation are both short-term and long-term in nature. Why is radiation hardening essential? In order to answer that we will present a brief case study of a recent satellite failure.
Figure 3: Phobos-Grunt failure featured in IEEE Spectrum February 2012
The Russian Phobos-Grunt satellite was supposed to put Russia back on the space map (after 15 years) and send China on its first inter-planetary mission. It was launched in November and was supposed to be a sample return mission from one of the Martian moons, Phobos. The mission cost was $64 million. The satellite went into orbit, failed it's first burn, and fell back down to Earth.
The official investigation blames the loss of the probe on memory chips that became fatally damaged by cosmic rays. The failure mechanism is speculated to be the simultaneous disabling of two identical chips in the dual-computer control system, causing both to restart simultaneously. A 512 kB SRAM used in the mission may have been corrupted by Single Event Latchup.
"He points out that neither the original fabricators nor the commercial vendors test for radiation, and they would not give radiation specs."
Long Term Radiation Effects
Total Ionizing Dose
Lower energy particles deposit charge in active areas of oxides in devices causing shifts in threshold voltages. Cumulative shifts over many devices and over long periods of operations cause leakage currents to flow in NMOS devices when devices should normally be off. Cumulative TID causes serious power draw issues on a spacecraft until failure.
TID is still measured in a deprecated unit known as rads. (SI unit is now the Sievert where 100 rad = 1 Sv). A lethal dose for a human is on the order of 10krad, while radiation hardened devices can withstand up to 1-10's of Mrad in Si.
Displacement Damage
Higher energy particles embed themselves in the lattice structure of a semiconductor and lead to recombination and deep-level traps. These recombination centers change the performance parameters and reliability of devices over time.
Displacement Damage dose is measured in Mev/g wrt (Si).
Short Term Radiation Effects
Single Event Effects
Single event transients are dynamic radiation events where pulses of radiation can cause currents to flow in devices and substrates anywhere in a system. SEE is observed in measurements in terms of energy deposited by the particle along the trajectory is known as the Linear Energy Transfer (LET) in terms of MeV*cm^2/mg. The higher the LET for a given cross section of a device, the more likely that there will be an effect on that device.
TID is still measured in a deprecated unit known as rads. (SI unit is now the Sievert where 100 rad = 1 Sv). A lethal dose for a human is on the order of 10krad, while radiation hardened devices can withstand up to 1-10's of Mrad in Si.
Single Event Upset
Depending on what is struck, this can cause single bit upsets in a system. In severe cases, multiple bit upsets can be triggered that cannot be recovered from by the system until a complete system reset is performed.
Single Event Latchup
Often due to a parasitic PNP device, event that changes the state of a device or circuit and induces a high current state. A current transient from a particle in this parasitic PNP can cause latch up from positive feedback. This can result in permanent damage if high enough currents are induced. Requires power cycling to reset the device.
Radiation Hardening
While the task of radiaiton hardening sensitive communication systems on a satellite can be daunting, a combination of the following methods are most typically used for the electronics in any given satellite.
Shielding
Encasing electronics in aluminum shielding with a thickness of up to 20 mm reduces both static and dynamic doses by orders of magnitude over the time period of a mission. However the tradeoff is in the additional weight of the spacecraft and deciding what electronics to move into electronics vaults. Some electronics may suffer hits to performance when moved away from other instrumentation, there will be additional cabling lengths and weight, and not all instruments can be relocated into the vault.
Radiation Hardening By Design
This is through using a radiation hardened process or by employing circuit design and layout techniques to bring radiation tolerance to a non-radiation hardened process. Process changes to reduce and/or eliminate latchup involve silicon on insulator (SOI), deep trench, and n+ rings; however, these add higher costs to a process due to additional masking steps.
Rad-hard COTS Parts
The idea is to save on hardware design costs by using commercial off the shelf parts. There is a reason why radiation hardened parts are expensive as radiation test time itself is extremely expensive.
Radiation sources are either ion beam, electron beam, x-ray, cobalt, etc. Sources are not combined, so the effects from a proton radiation measurement will differ from an electron beam radiation measurement. Low dose rate cobalt sources can takes months of testing to get dose rates that are representative to what the electronics will experience in operation. Thus higher dose rate sources are typically used to accelerate testing to hours worth of time.
SEU Mitigation
As an energetic charged particle passes through a semiconductor material it deposits energy in the form of ionized charge (electron-hole pairs) along its trajectory, forming a dense track of mobile charge pairs.
Figure 4: Ion Track through a cross-section of semiconductor material
CMOS and SiGe HBT technologies are not immune to the effects of SET. Heavy ion strikes to sensitive junctions will cause currents to flow regardless of process changes. Designs must be tolerant to SEU at higher system levels of design through redundancy, error detection and correction, and shielding.
The variety of coding methods already in use will help to prevent catastrophic failure during a single event upset:
Parity - detect only code that does not correct bit upsets itself
Hamming - correct single bit upsets
Reed-Solomon - block error codes, can correct up to 16 consecutive bytes in error
Convolutional - interleaved, can correct burst noise
Vibration Testing Procedure at JPL
IEEE Spectrum: Phobos-Grunt Failure
Avery, K., "Radiation Effects Point of View", 2009 IEEE Nuclear and Space Radiation Effects Conference Short Course Notebook
Phillips, S. D., "Single Event Effects and Radiation Hardening Methodologies in SiGe HBTs for Extreme Environment Applications", Ph.D. dissertation, Georgia Institute of Technology, 2012.
Arora, R. A., "Trade-Offs Between Performance and Reliability of Sub 100-nm RF-CMOS Technologies", Ph.D. dissertation, Georgia Institute of Technology, 2012.