One of the most interesting growth applications to emerge in the field of embedded controls in recent years is the small satellite. While 'terrestrially-focused' embedded engineers have been riding the IoT wave, a growing number of engineers have been pitting their wits against the forces of the galaxy by designing embedded systems that operate in small satellites orbiting the earth.
There are a few different categories of small satellite types based on weight, but the most popular is the CubeSat. The CubeSat standard has its roots in academia and has given rise to an ecosystem of components that can be connected together relatively easily to create a working satellite. A 1U CubeSat has a form factor of 10 x 10 x 11.35cm. The CubeSat chassis is shown in Figure 1 along with a standard form factor circuit board that fits into the chassis. These can be concatenated with additional 1U cubes to form larger systems such as 3U or 6U depending on the complexity of the satellite system requirements.
Figure 1. 1U CubeSat chassis and standard-sized controller board (Source: VORAGO Technologies)
The CubeSat standard has proven to be extremely successful. By 2014, approximately half of the satellites that were launched into orbit were CubeSats. The expectation is that most of the working satellites in orbit will soon be CubeSat types. Also in 2014, the number of commercial CubeSats overtook the number from academia for the first time. Compared to conventional satellites, CubeSats are inexpensive to develop and launch, and can be networked together to form constellations. Such a configuration has inherent redundancy and potentially greater coverage when in orbit. These benefits are appealing to commercial satellite developers, who have embraced the standard.
CubeSats are now used by academia for research, government entities and the military for science purposes, and commercial companies for applications such as telecommunications, video, and sensing applications.
This new growth area in satellite technology is not without its challenges, however. Many CubeSat missions fail. Based on data going back to 2000, over 40% of CubeSat missions were categorized as launch fail, DOA, or early loss. There have been many reasons cited for this high failure rate, such as ambitious technology infusion and the lack of testing, possibly related to low budgets in the hobbyist and academic sectors. It does appear that you get what you pay for.
The challenge facing all embedded designers who are developing systems that operate in space is the limited pool of components that were designed to operate in this environment. Semiconductors that operate in space need to be 'hardened' to withstand the effects of radiation. There is not a huge portfolio of components to choose from, and those that are available are not exactly leading edge or 'state-of-the-art' and they tend to be very expensive. Initially, the CubeSat community threw caution to the wind and used commercial off the shelf (COTS) products in their designs. This is one of the reasons why that there was a large failure rate in those missions.
While the higher-end government agency funded CubeSat developers have always opted for radiation-hardened components, there is now a trend to also opt for rad-hard devices within the entire CubeSat community. Even low-budget CubeSat developments are likely to cost upwards of $100,000. The new mindset is that the incremental cost of a rad-hardened microcontroller at the heart of the system is a good trade-off and doesn't impact the overall budget in a big way.
As the majority of CubeSat missions have transitioned away from academic and hobbyist projects to commercial ventures, expectations about the specifications and reliability of the CubeSat have also changed. Mission lifetimes have been extended from months to years, and the orbits in which these missions will fly have also been extended. This exposes the CubeSats to more radiation where they are at risk of malfunction through Single Event Upsets (SEU) from ionizing particle strikes or a build-up of Total Ionizing Dose (TID) that will negatively affect the CMOS devices.
An SEU that is caused by a particle strike can upset memory bits, upset logic operation, or cause latch-up. Latch-up occurs when parasitic devices on the die switch-on and create a short circuit from Vdd to Vss. This condition is often irrecoverable and will destroy the device. TID accumulates over time and results in increased source-drain leakage as the oxide builds up an accumulated charge. There is also an expansion of the depletion region between PMOS and NMOS type devices (this becomes even more of an issue as semiconductors shrink further down the process geometry curve). TID accumulation will result in increased leakage current, and eventually the CMOS device will cease functioning as the supply voltage is pulled down.
An ionizing particle strike on a CMOS die is shown in Figure 2. High energy radiation creates many electron-hole pairs when striking the semiconductor structure. These electron-hole pairs create current and voltage spikes. A Buried Guard Ring (BGR) structure has been implemented on the CMOS structure in Figure 2 to protect the CMOS device by stopping the parasitic bipolar devices from switching on and triggering latch-up.
Figure 2. An ionizing particle strike into a CMOS die can cause latch-up if parasitic bipolar devices are triggered (Source: VORAGO Technologies)
Radiation effects on CubeSats are accelerated as they do not have as much shielding on the chassis as do larger satellites. Normally, CubeSat chassis panels have solar cells mounted on them.
As low-budget CubeSats evolve from using strictly off-the-shelf components to becoming more radiation tolerant, it is still possible to optimize the system for cost. A popular approach is to use a rad-hard MCU as a safety monitor device in the system, and to use commercial grade components elsewhere in the system where they are less critical and can be monitored by the rad-hard MCU. An example is given in Figure 3. The VA10820 MCU is used to boot-up and configure the Virtex-5 FPGA. After the FPGA has been loaded and is operational, the rad-hard MCU (which is itself latch-up immune) will continue to monitor the operation of the FPGA. In the event of an FPGA error caused by radiation, the MCU can reconfigure the FPGA or even reset it if latch-up occurs. This approach keeps within the spirit of a low-cost approach to CubeSats, but is a solid solution to the problem of space radiation. (Note that rad-hard FPGAs are expensive — very high-end automobile level of expensive!)
The same rad-hard MCU can be used to monitor other parts of the system that do not have as high a degree of tolerance for radiation. The CubeSat standard uses a PC/104 bus system, which allows different boards to be stacked inside the CubeSat chassis and communicate through a common bus structure. It is possible to buy off-the-shelf solar cell arrays, battery pack modules, power management modules, communications modules, and attitude determination and control system (ADCS) modules from different CubeSat component suppliers. The payload is the customized board that performs the core mission of the satellite. This could involve data collection and will likely utilize various sensors. The payload board communicates on the same PC/104 bus as the other CubeSat boards, and it will share the same resources such as access to power.
Figure 3. Using a rad-hard MCU with a COTS FPGA to optimize cost and mitigate radiation effects (Source: VORAGO Technologies)
CubeSats have opened-up a lot of opportunities for space science and exploration. From hobbyist science projects, the standard has evolved into a platform for deep-space exploration in only a few short years. The early polarization of either “very expensive rad-hard everything” or “very cheap commercial-off-the-shelf everything” is settling to a natural equilibrium point in which non-rad hard components can sometimes be used, provided they can be monitored and controlled by a rad-hard anchor chip such as an MCU.
The number of CubeSats that will be launched are expected to reach around 500 per year by the end of this decade. There is more private enterprise investment in space than ever before, and more embedded electronics designers working on space projects than ever before. When you get bored of the Internet of Things, consider getting involved in enabling the Internet of Space!
Ross Bannatyne currently serves as marketing director for VORAGO Technologies, a privately held, fabless semiconductor company based in Austin, Texas. Ross is a veteran of the semiconductor business having worked in various engineering, marketing, and management roles in Europe, Silicon Valley, and Texas. Ross is the author of two textbooks — Using Microprocessors and Microcomputers (Prentice Hall) and Electronic Control Systems (Society of Automotive Engineers) — and holds patents in failsafe electronics and microcontroller development tools. When he is not working, Ross relaxes by watching Glasgow Celtic Football Club and the Oakland Athletics.