Quad Magnetometer Array Redesign

Overview

The Space Physics Research Lab (SPRL) at the University of Michigan, in conjunction with the private company General Orbit, successfully developed and tested a novel magnetometer design which utilized four spatially separated tri-axis arrays separated such that ambient fields acting upon the system could be isolated and accounted for.

The most significant advantage of such a design is that it could be placed within a spacecraft (assuming no high-intensity EMF from e.g., a motor is present), whereas typical magnetometer systems must be deployed multiple meters away from the main craft body via an extendable boom to avoid magnetic noise or interference from the craft structure itself. The design built off of prior work in magneto-inductive sensing and multi-sensor magnetometer arrays (Regoli et al., 2018; Strabel et al., 2022), where low-cost COTS sensors were shown to achieve sub-10 nT resolution and where multi-sensor configurations enabled resolution improvements through spatial oversampling.

Despite the success of these models, they were not flight-ready or tested. In 2026, SPRL was selected to fly a quad-mag system on the upcoming NASA-run Resolute mission, a double sounding rocket flight with the primary goal of characterizing the flow of heavy ions from the ionosphere into the magnetosphere. I was tasked with redesigning the system, not only to ensure SPRL could successfully take part in this mission, but also making the system robust enough that it could be used on other future flights. I focused on the following key improvements:

• Maximizing board outline to push spatial separation to physical limits constrained by physical housing
• Identify and replace consumer-grade components (passives, ICs, microcontrollers, etc.) with mil-spec/aerospace components
• Redesign trace layout and component placement to keep varying signals (primarily communication) as far away from magnetometers as possible
• Create a low-noise power topology capable of converting up to 36 VDC input down to 3.3 VDC.
• Reduction in overall power usage

Hardware Design

---

Overall Architecture

The system is designed to be serviced by a single connection point to the rest of a parent spacecraft. This interface handles both power and data operations, communicating over the RS-485 protocol. A single, robust LDO drops the input voltage down to 5 VDC, and this voltage is then further dropped to 3.3 VDC via a distributed array of smaller SOT23 LDOs. Eight of these directly service the discrete analog and digital input lines for each of the four sensor arrays; the supplied documentation recommends a discrete supply for each. The final LDO powers the primary microcontroller and a temperature probe, used for calibration and as an emergency thermal shutoff.

Power

Usage of a space qualified version of the low-power TI MSP430 series of microcontrollers brought total board current draw from the 5 V LDO down to only 13 mA. Unfortunately, in the worst case at 36 V input, this still leads to ~450 mW of power dissipated as heat in the initial LDO even at this low current level. Both buck-converter and charge-pump (Dickson converter) designs were examined in an attempt to introduce a power conversion stage with higher efficiency, however due to the nT-range sensitivity of the measurement suite this proved difficult to realize. Most switching-mode converters rely upon an inductor to both facilitate voltage conversion as well as power smoothing, however this also radiates significant EMF from the component itself. Even when considering advanced, shielded inductors, simulation showed that the best designed switching converters would still introduce potentially hundreds to thousands of nT of magnetic field variance when placed within 5 cm of the magnetometers. Even an inductorless charge-pump topology still creates a significant magnetic field in the hundreds of nT range, simply from current travelling through PCB traces. In both cases, these errant fields are not constant as they are load-dependent, and cannot be easily accounted for by exploiting the spatial separation effect discussed above. The only robust solution to this issue is to have a separate, isolated power conversion stage in a different physical location from the rest of the system, a luxury not often afforded in spacecraft design. As such, for the current prototype designed for the Resolute mission, I decided to simply accept the poor power conversion efficiency, as the maximum possible power draw of the system including waste heat was only 2 W.

Communication

The system was equipped with an RS-485 transceiver, capable of both receiving instructions/commands as well as sending data. RS-485 is an evolution of RS-422, capable of having multiple sending and receiving nodes upon the same network. Furthermore, it is only a two-wire protocol compared with RS-422's four-wire requirement (for bidirectional communication). Both still use differential signalling, and are long-term evolutions of the serial RS-232 protocol.

RS-485 has another advantage which led to its selection for this project: it is fully compatible with an RS-422 network assuming the device only needs to send or receive, not both. As such, an RS-485 transceiver allows for bidirectional communication on compatible systems, but can still send data on more complex networks (such as Resolute's), with the small downside of being unable to directly control the sensor package actively. As mentioned previously, total system power draw can push 2  watts; this jump is almost exclusively due to switching losses when this transceiver is actively communicating. This particular transceiver, however, was configured to operate at a maximum of 100 kHz, but the sensor sample rate is only 50 Hz. Doubling this value to provide headroom, under normal data sending operations the transceiver only needs to operate at a 0.1% duty cycle, so nominal power draw for this subsystem is still considered fairly low.

Final Prototype Design

Below is the final design, posted with permission from but blurred at the request of General Orbit and SPRL.

High-speed and high-power components were kept centrally located but slightly offset from the magnetometer array. Increased distance reduces EMI/EMC issues, however, placing components or traces near the exact geometric center increases common mode coupling, hindering the ability to calculate and offset noise via the spatial gradient.

This design has passed PDR/CDR with NASA, and final internal reviews are currently underway as of Spring 2026. The design will be fabricated and tested electrically, thermally, and mechanically (e.g. vibration and stress testing) in Summer 2026. This system is slated to fly on NASA's Resolute mission in 2027.