CubeSat Solar X-Ray Flare Payload — Rev A

The satellite's objective is to detect and characterise solar flares by measuring changes in solar soft-X-ray irradiance. It would observe the Sun whenever orbital and attitude conditions permit, measure X-ray intensity, and identify the beginning, peak and end of an event, recording how quickly intensity rises and falls and estimating flare strength from calibrated measurements.

Supporting measurements attach context: visible light for Sun-presence and pointing checks, local magnetic field as space-environment data, and time, position, attitude and temperature on every record. Measurements are stored until they can be transmitted, and compared afterwards with established space-weather observations.

The payload does not predict a flare. It detects and characterises one after its X-ray output begins increasing.

A flare appears first as a rise in soft X-rays above a slowly varying background, so the instrument watches one band continuously and looks for departures from that background rather than for an absolute level.

A possible initial measurement band is approximately 0.1 to 0.8 nm, which is the band conventionally used for solar-flare classification. This is a starting target, not a specification: the real spectral range will be set by whichever detector and filter combination is eventually chosen.

Onset detection is a change detector, not a threshold on absolute intensity. The software maintains a running estimate of the background X-ray level, and a trigger requires both a minimum increase above that background and a minimum persistence time, so a single high sample cannot start an event.

The trigger settings are adjustable rather than fixed, and eclipse transitions and loss of pointing must not produce a flare alert — both look like a large, fast change in the raw series.

The targets to design against: flare mode begins within 10 seconds of the trigger condition being met, at least 95 per cent of validation events above the selected threshold are detected, and false alarms stay below the mission limit. These are acceptance criteria for a future test campaign, not results.

Once an event is open the software records the largest valid X-ray value, and keeps the event open until the signal has stayed below an ending threshold for a defined time. Closing on the first sample below threshold would split one flare into several.

Multiple peaks within an event are retained rather than collapsed, because a flare with two maxima is a different observation from two flares. Each event record carries onset, peak, end, duration and maximum intensity.

Validation is by replaying single-peak and multi-peak profiles, slow and rapid decays, and interrupted observations, and comparing against independent analysis software.

Two independent things have to be true before a measurement counts: the Sun must be present, and the detector must be pointed at it.

Presence comes from the VEML7700, kept inside its measurement range with optical attenuation if direct sunlight would saturate it, and synchronised with the X-ray samples. Its output is a quality check. Visible light alone never constitutes flare evidence.

Pointing comes from the spacecraft. The detector axis is measured relative to the spacecraft coordinate frame, the ADCS supplies timestamped attitude, and a measurement is valid only when the Sun is inside the detector field of view, with attitude uncertainty carried into the quality information. Off-axis measurements are marked invalid or reduced quality rather than discarded silently.

The magnetometer records the local three-axis field, synchronised with the flare observations. Calibration coefficients correct hard-iron and soft-iron effects, and out-of-range readings are flagged.

The harder problem is the payload itself. The MCU, the microSD card, the LEDs, the regulator and any radio all generate magnetic fields, so payload-generated interference has to be measured rather than assumed small, and the sensor placed away from magnetic hardware and large current loops. Planned verification runs a Helmholtz coil and known rotations against an independent magnetometer with all of that hardware active, because testing a quiet board proves nothing about a working one.

Detector gain and offset move with temperature, and in orbit the temperature moves a great deal. A sensor sits near the X-ray detector, gain and offset are characterised across temperature, and the correction coefficients are stored in memory. Measurements taken outside the calibrated temperature range are flagged rather than corrected with coefficients that do not apply.

The targets: temperature known to within 1 degree Celsius, corrected readings repeatable within 5 per cent for identical inputs, and calibration drift below 10 per cent after environmental testing. These are the limits the design has to meet, established by thermal-chamber testing that has not been carried out.

A flare measurement that cannot be placed in time cannot be compared with anyone else's. Absolute time comes from the spacecraft, with a target accuracy of 100 milliseconds or better, and every record carries spacecraft-position information alongside it.

Clock drift is characterised rather than assumed, and loss of synchronisation is flagged. The rule that follows is simple: no accepted data packet contains an unknown time.

Science and housekeeping data are written to the microSD card until they can be downlinked. A record holds X-ray intensity, light level, magnetic field, temperature, time, position, attitude, operating mode and quality flags, with a checksum or CRC on each packet.

Flare data takes storage priority, and critical data is also buffered in internal flash where practical, so a card failure does not take the event with it. The design constraints are about failure rather than capacity: an interrupted write must not corrupt unrelated files, and a full card must not crash the payload.

The targets are at least 99 per cent of test records recovered correctly, corrupted data detected rather than returned, and a sudden power loss that does not destroy the whole dataset.

The payload talks to the spacecraft computer, not to the ground. A method is still to be selected — UART, SPI, I2C or CAN — and commands cover modes, thresholds, calibration, data download, timing and reset. Packets carry sequence numbers and error detection, invalid commands are rejected safely rather than partially applied, and the link recovers automatically after an interruption.

One point is worth stating plainly, because the schematic invites the opposite assumption: the MCU's 2.4 GHz radio does not serve as the satellite downlink. The payload sends its data to the spacecraft computer, which controls a dedicated space-qualified radio.

The AP2112K produces the 3.3 V rail, and its input voltage has to stay above the level it needs to regulate — a Rev B verification item, not something Rev A confirms. The polyfuse limits excessive current and the USBLC6-2SC6 protects the USB data lines against electrostatic discharge.

The requirement that matters for the science is quieter: sensitive analog circuits need properly filtered power, because the eventual detector front end will be measuring a very small current while the microSD card and the radio are switching. Proposed consumption is below 2 W nominal and 3 W peak, and the intended checks include that microSD operation does not reset the processor and that power noise does not interfere with X-ray measurements.

The aim is that no software or hardware problem permanently disables the payload. A watchdog resets frozen software, sensor communication timeouts are detected, invalid sensor readings are rejected rather than stored as data, and fault information is written to nonvolatile memory so it survives the reset that follows.

Repeated faults put the payload into safe mode, and the spacecraft can reset or power-cycle it. The condition that has to be designed against explicitly is the endless reset loop: recovering from a fault is only useful if the recovery itself terminates.

The payload has to survive launch and then work in orbit. The PCB and detector assembly must tolerate launcher vibration and shock, operate in vacuum, and use materials that meet outgassing limits, with every component assessed for radiation effects.

The requirement specific to this instrument is alignment. The X-ray filter and the detector alignment have to remain stable, because a detector that survives launch but is no longer pointed where the calibration assumed is not a working detector. Planned verification is vibration and shock, thermal-vacuum cycling, electromagnetic-compatibility testing and radiation analysis, with calibration repeated afterwards to measure what the testing changed.

Calibration links a number the instrument produces to a physical quantity, and without it the measurements are only relative.

The intended approach is calibrated X-ray equipment at a licensed laboratory, sweeping intensity from the minimum to the maximum of the required range, repeated at several temperatures, with a separate measurement taken with the aperture blocked to characterise the dark background. The resulting correction table is validated against separate test measurements rather than against the data used to build it.

Science data is then expected to be comparable with established solar-monitoring spacecraft, allowing for the difference in spectral range between instruments.

The review compared the schematic against the eighteen mission functions and the fourteen top-level requirements, and asked of each one whether Rev A could contribute to it at all.

What is present: the payload computer, the 3.3 V regulation chain, input and ESD protection, microSD storage, the two context sensors, a comparator, the USB-C and Qwiic development interfaces, the expansion and payload-bus headers, and the buttons and LEDs. That covers the housekeeping half of the instrument and a useful amount of the context half.

What is absent is the measurement. There is no detector, no aperture, no filter, no low-noise front end and no ADC path, which means TR-01 through TR-05 — the requirements that make this a science instrument rather than a development board — have no hardware behind them. The LMV331 is labelled a radiation pulse comparator, but with no conditioned signal reaching its input and no defined threshold at the other, it cannot perform that function as drawn.

The review also flagged several items that are not missing subsystems but unfinished connections: an absent pull-up on the comparator output, unverified regulator input voltage, magnetometer placement close to current-carrying electronics, unlabelled payload-bus connections, and no test points on the analog path that does not exist yet.