An isolated CAN bus can help small satellite designers connect distributed electronics while separating sensitive controller circuits from the bus-side environment. As satellite platforms become smaller and more cost-constrained, serial networking offers a way to reduce wiring and board area. But isolation only works as intended when designers account for both the data path and the power domains around it.
CAN, or Controller Area Network, is a two-wire serial bus used to exchange information among controllers and other processing devices. Its differential signaling, multi-node capability and relatively straightforward interface make it relevant to satellite systems that need to link sensors, processors and other equipment across a payload or spacecraft.
Why CAN bus suits small satellite networks
Small satellites are typically defined as having a mass below 500 kg. Compared with traditional spacecraft, they face tighter constraints on size, weight and cost. These limits can make it difficult to use conventional space-qualified components throughout a design, even when some level of quality or radiation tolerance is required.
CAN addresses some system-level constraints by carrying communications over a two-wire differential pair rather than a parallel set of signal lines. A single bus can connect multiple nodes, and destination information travels in the messages themselves, avoiding a separate select line for each device. This can be useful when several processing units need to exchange data without dedicating a point-to-point interface to every connection.
In a satellite payload, for example, sensors and positioning devices may be distributed across a board. CAN provides a shared link for those nodes to report data to other parts of the system. The source whitepaper describes a network supporting up to 30 nodes and a maximum data rate of 1 Mbps; actual network design must follow the relevant device and system requirements.
Understanding CAN bus layers and signaling
A CAN node generally comprises a processing unit, a CAN controller and a transceiver. The processing unit may be a microcontroller, FPGA or microprocessor. The controller formats and manages messages according to the CAN protocol, while the transceiver converts between controller-level signals and the electrical levels used on the bus.
The protocol separates responsibilities across layers. The application layer handles messages and status information, including filtering. The transfer layer manages framing, bit timing, synchronization, arbitration and error detection. The physical layer defines the electrical interface, including the differential pair commonly identified as CANH and CANL.
Because CAN is a shared bus, nodes must coordinate their transmission and sampling of bits even though the network does not use a master clock. The physical network also requires termination: a 120-ohm resistor is placed differentially between CANH and CANL at each end of the bus. Correct termination is important to achieving the intended signal levels.
What isolation protects—and what it does not
Isolation can help protect controller electronics from high-voltage transients and prevent ground loops between the controller side and the CAN bus side. In a space application, the bus may also be exposed to transients associated with single-event effects from heavy ions. An isolation barrier in the data path helps keep disturbances on one side from directly coupling into the other.
However, a digital isolator alone does not create a fully isolated node. The controller and isolator require a controller-side supply, while the transceiver operates from the bus-side supply. Those power domains must also remain isolated. If the grounds or supplies are tied together across the barrier, the intended separation is undermined.
For a typical isolated arrangement, the digital isolator sits between the CAN controller and the CAN transceiver. Transmit data passes from the controller through the isolator to the transceiver; received data travels back across the barrier to the controller. The controller-side ground and supply remain separate from the bus-side ground and supply.
Radiation-tolerant components and compact implementation
The whitepaper describes pairing Renesas’ ISL71710M active digital isolator with the ISL71026M CAN transceiver to build an isolated CAN node. The ISL71026M is a 3.3 V transceiver rated for CAN data rates up to 1 Mbps. The ISL71710M uses giant magnetoresistive (GMR) technology: an input-driven coil changes the resistance of GMR elements across the isolation barrier, which changes the output state.
The components are presented as part of a radiation-tolerant plastic product approach for small satellite applications. The whitepaper notes that radiation data is published in the relevant datasheets. Designers should review that data and the devices’ qualification information against mission requirements rather than assuming that a component’s radiation-tolerant designation alone makes it suitable for every orbit or mission profile.
Package area and cost are also design considerations. The paper’s comparison lists package areas for selected transceivers and isolators, illustrating that available implementations differ substantially. Plastic packaging and a production flow without some of the tests used in standard Class V processing are described as ways to reduce cost. Those tradeoffs should be evaluated alongside mission assurance, screening and radiation requirements.
Practical design checks for an isolated CAN node
- Define the network: Identify the number and location of nodes, the required data rate and which devices share the bus.
- Plan termination: Place the specified differential termination at both ends of the bus.
- Keep power domains separate: Maintain isolation between controller-side and bus-side supplies and grounds.
- Check radiation data: Compare published component data and qualification details with the mission’s requirements.
- Assess physical constraints: Consider board area and cost without overlooking system-level reliability needs.
Conclusion
An isolated CAN bus can provide a compact way to connect multiple satellite electronics nodes while helping shield sensitive controllers from bus-side transients and ground-loop effects. The central design lesson is to treat isolation as a system property: the data signals and their power domains must both respect the barrier. With network topology, termination, radiation data and packaging tradeoffs addressed together, engineers can assess whether this approach fits the size, cost and protection requirements of a small satellite.