Radiation testing of COTS data-acquisition electronics for space applicationsStory
June 21, 2017
Thanks to its proven reliability, cost-effectiveness and size, weight, and power (SWaP) optimization, commercial off-the-shelf (COTS) equipment is becoming increasingly attractive to system designers for use in space applications. As COTS equipment extends into mission-critical functions, the risks to the COTS electronics posed by space radiation need to be understood and mitigated.
Radiation is the emission of energy as electromagnetic waves or as moving subatomic particles. We encounter many types of radiation in our everyday lives including sound, light, radio waves, wireless signals, etc. While these forms of radiation are generally low energy and benign, there are some forms of higher energy radiation that can cause damage to materials through the process of ionization. Ionizing radiation can produce charged particles in matter through which the radiation passes. These charged particles pose a threat to semiconductor material.
While there are very few natural sources of ionizing radiation on Earth, there are two main sources of ionizing radiation in near-Earth space. The main source of radiation near Earth is the sun. The second main source of radiation in near-Earth space is galactic cosmic rays (GCR), which emanate from outside our solar system. In space, ionizing radiation typically takes the form of highly energized particles consisting of protons, electrons, and heavy ions. The energy levels of these particles can range in intensity from kilo-electron-volts up to giga-electron-volts. Protons, originating from solar activity and GCRs, are trapped in the Van Allen radiation belts – doughnut- or torus-shaped regions of particles trapped in the Earth’s magnetic field. Electrons are also found as trapped particles in the Van Allen belts. Heavy ions, which also originate from solar activity and GCRs, are essentially atomic nuclei with the electrons stripped off. They occur in space with a distribution similar to the natural distribution of elements on Earth.
Solar activity, such as solar flares and coronal mass ejections, can increase the amount of radiation near Earth and affect the amount of radiation encountered by a space vehicle near Earth. On the other hand, solar winds – streams of charged plasma emanating from the upper atmosphere of the sun – can shield the Earth from GCRs and reduce the amount of radiation near Earth.
There is typically an 11-year cycle between maximum and minimum solar activity: Seven years of maximum solar activity can increase trapped proton populations in the Van Allen belts, but shield the Earth from GCRs, followed by four years of solar minimum activity, which can reduce the trapped proton populations in the Van Allen belts, but also brings reduced shielding from GCRs. (Figure 1.)
The other major influence that can affect how much radiation a space vehicle encounters is the vehicle’s orbit or trajectory. The vehicle’s trajectory in relation to the Van Allen belts and the Earth’s magnetic fields can have a significant effect on the intensity of radiation it experiences.
Figure 1: Artist’s rendering of magnetic fields and Van Allen belts. (Image courtesy of NASA.)
Radiation’s effects on semiconductor material
When an integrated circuit (IC) is penetrated by an ionizing particle, the particle’s energy is deposited in the IC’s semiconducting material. When a large number of ionizing particles pass through an IC’s semiconductor material over time, the deposited energy can damage the crystalline structure and affect the performance of that semiconductor material. This accumulated damage is the total ionizing dose (TID) and measured in Rads or Grays (Gy).
If a single ionizing particle passes through semiconductor material and deposits enough energy in an active area of the circuit, the result can be an immediate and undesirable effect on the operation of the IC, called a single event effect (SEE). SEEs that can occur in modern electronic components are single event upsets (SEU), single event latchups (SEL), single event transients (SET), single event burnouts (SEB), and single event gate ruptures (SEGR).
An SEU is the unintentional change in the state of a memory cell. For example, an SEU can cause a bit-flip, or a single bit in RAM changing its value from “0” to “1”. After an SEU, the incorrect bit value may or may not affect the operation of a unit depending on what it is being used for. The value in the bit can be reset again through normal operation.
An SEL occurs when an ionizing particle creates an unexpected path in the semiconductor material of the IC that enables current to flow. This self-maintaining current path is created by a parasitic bipolar transistor being turned on and manifests as an increase in current drawn by the circuit. It may or may not affect the functionality of the circuit. This increased current can lead to excessive heat dissipation in an electronic component, and can result in permanent damage if the SEL is not mitigated by power-cycling the IC.
The question then arises: What can or should be done to mitigate against SEEs? Generally speaking, a system requirement for operation in space is a particular level of reliability that ensures the system will actually operate in a radiation-filled environment. Historically, systems that were designed for use in space took the radiation environment into account during the design phase. However, this is not always the case today when modifying COTS equipment for use in a space radiation environment. First, the system integrator should determine whether the radiation environment requirements are met using the equipment as is. If not, reliability can be increased using redundancy or extra shielding. However, this approach will result in a mass gain that may break other requirements.
Case study: testing a COTS data-acquisition system
An example of COTS hardware that has been qualified for operation in a space radiation environment is Curtiss-Wright’s Acra KAM-500 data acquisition unit (DAU). Testing was carried out on a number of different KAM-500 modules used in the DAU to characterize their susceptibility to both SEU and SEL-type upsets. The KAD/ADC/141/PT100 KAM-500 module is a 16-channel RTD type sensor module that provides constant current excitation and monitors voltage across PT100 sensors.
In May of 2016, this module was tested at the Paul Scherrer Institut (PSI) in Switzerland. (Figure 2.) During the test, the module’s entire double-sided PCB was exposed to a 200 MeV proton beam while operational. The currents of the voltages used to power the module were monitored and the module was automatically power-cycled if any of the currents exceeded a threshold. Every time the module was power-cycled was considered a SEL type event. In addition, the readings the module took of the attached PT100 sensors (simulated using precision 100 Ohm resistors) were monitored for accuracy. If one or more channels showed a value outside of the expected accuracy, that event was considered an SEU.
Figure 2: Example of proton test setup at PSI. Photo: Curtiss-Wright.
The module was exposed to a proton beam with an average flux of 3.09E+6 protons/cm2·s up to a fluence of 1E+10 protons/cm2. During this exposure, there were 20 SEL-type events and zero SEU-type events detected.
Calculating on-orbit error rates for SEL-type events
The following examples illustrate how the Bendel A and Figure of Merit (FOM) methods were used to estimate on-orbit SEE event rates.
- Proton induced upset rate example
The Bendel A method  was used to estimate the proton induced upset rate for SELs in the KAD/ADC/141/PT100 module at the nominal International Space Station (ISS) orbit. In order to use CREME-96 (a tool to predict SEE rate) to estimate the upset rate, the A parameter must be determined. This can be done by iteratively solving the Bendel A formula until a value for A is found that satisfies the experimental results that were seen, e.g., 20 events. To do this, first the cross-section for 20 events was calculated for a fluence of 1E+10 protons/cm2. Next, the cross-section was calculated using the Bendel A formula with different values of A until a cross-section equal to the experimental cross-section was achieved. This was done using a script, the output of which indicates that an A value of 13.6511854 MeV results in a cross-section of 2.000E-9 cm2. Next, CREME-96 was used to calculate the expected proton induced upset rate using the Bendel A model of the KAD/ADC/141/PT100 module for the ISS orbit.
The output of the PUP (Proton Upset) module in CREME-96 showed that the KAD/ADC/141/PT100 module can expect to see approximately 2.97040E-03 SEL events per day, or one event every 336.65 days. These are proton induced upsets and based on the ISS orbit with a shielding of 118 mils of aluminum.
- Heavy ion induced upset rate example
The FOM method  was used to estimate the heavy ion induced upset rate for SELs in the KAD/ADC/141/PT100 module at the nominal ISS orbit. First, the FOM for the KAD/ADC/141/PT100 module was calculated by assuming the 200 MeV test results represented the limiting proton cross-section for the device. Next, the C value for the ISS orbit needed to be calculated using existing component test data. The resulting data showed that the KAD/ADC/141/PT100 module should experience approximately one SEL type event every 353.44 days due to heavy ions in the ISS orbit.
Understanding the risks posed by the environment is important when using COTS equipment in space. Mitigation strategies can then be formulated that best meet the reliability needs of the mission without negatively affecting other considerations, for example, adding significant weight to the systems.
One mitigation technique against SEL events is to monitor the current draw on voltage rails to detect spikes and then power-cycle only the affected circuitry. A technique used for SEU events is to continuously refresh all RAM in the system such that any errors are flushed out quickly.
It is possible to characterize existing COTS electronics for the low Earth orbit (LEO) space radiation environment through a single high-energy proton test at the PCB or system level. The results of these tests can guide the space vehicle system designer in determining the need for mitigation techniques to increase the reliability of a COTS-based system through careful system design or unit level redesign.
The previous example is one in which testing was undertaken on a COTS module to determine an on-orbit SEE event rate using the Bendel A and FOM analysis techniques. The data garnered from these analyses can then be used to calculate what mitigation techniques, if any, are required to meet a minimum level of reliability in radiation-prone environments.
 W. L. Bendel and E. L. Petersen. “Proton Upsets in Orbit.”IEEE Transactions on Nuclear Science 30.6 (1983).
 E. L. Petersen. “The SEU Figure of Merit and Proton Upset Rate Calculations.” IEEE Transactions on Nuclear Science 45.6 (1998).
David Chamberlain is a systems engineer for Curtiss-Wright Defense Solutions in Dublin, Ireland. He holds a Bachelor of Applied Sciences degree in Electronic Systems Engineering from the University of Regina (Saskatchewan, Canada) and has worked at Curtiss-Wright since 2011. David’s other areas include systems integration and testing, customer training, and writing reams of documentation. Prior to joining Curtiss-Wright, David spent 5 years as an Applications Engineer at Xilinx.
Curtiss-Wright Defense Solutions www.curtisswrightds.com