CubeSat Storm-Time Radiation Belt Mapper — Rev A

The intended satellite would detect energetic charged-particle events, measure count rate along its orbit, compare channels behind different shielding, and attach time, position, attitude, magnetic field and temperature to every measurement. It would store the result, return it through the spacecraft, and let ground software build radiation-intensity maps along the trajectory.

One limitation belongs at the top, because the word "mapper" invites the wrong reading. A single satellite produces an along-track map. It does not produce an instantaneous map of the entire radiation belt; global coverage would need repeated orbital sampling, several spacecraft, or both.

Rev A is a schematic and a review. It demonstrates none of this.

A radiation measurement on its own is a count. What makes it a map is everything attached to it.

Each accepted interval carries a radiation level, the time, the spacecraft's position, its attitude and therefore where the detectors were looking, the local magnetic field, the temperature, and quality information saying whether the instrument was saturated, mid mode-change, or missing context. As the spacecraft moves, those samples line up into a profile of what it passed through.

The map is a record of what the spacecraft encountered along its trajectory, not a snapshot of the entire radiation belt at once.

Four detectors are not four times as informative as one. Four identical detectors behind identical shielding, looking the same direction, measure the same thing four times.

They become four science channels when they differ in ways that are known and deliberate: absorber thickness, aperture size, look direction, shielding arrangement, and the response that follows. A thicker absorber generally suppresses lower-energy particles more strongly, so comparing calibrated channels with deliberately different absorber responses can provide coarse penetration information.

That is the honest limit. The source puts it plainly: without shielding and calibration, four photodiodes cannot reliably determine particle energy or type. REQ 2 asks for at least coarse particle-penetration information, not species identification — and even that depends on geometry and absorbers Rev A does not define.

The shielding is not a box the electronics sit in. It is the part of the instrument that decides what each detector is measuring.

The source lists five jobs for it: block visible light, define the field of view, use different absorber thicknesses for different channels, prevent particles entering from uncontrolled directions, and provide a known physical geometry for calibration. A BPW34 responds to visible photons enthusiastically, and an aperture sets the solid angle a channel accepts.

That last job is why this is a measurement problem rather than a mechanical one. The geometric factor reflects the effective detector area and solid angle accepted through the aperture and shielding, and it enters the conversion from count rate to particle flux. Change the shielding and the number changes, though the environment did not.

None of it is on the sheet, and no absorber material or thickness is proposed here, because the source proposes none.

Two different questions, two different measurements, and they are not interchangeable.

Coincidence asks whether more than one channel responded within a defined timing window. A particle energetic enough to pass through one detector into another produces near-simultaneous events; a low-energy particle stopping in the first does not. So coincidence helps separate penetrating events from single-channel events, and both from noise and interference. It does not, on its own, identify what kind of particle arrived.

Pulse height asks how much energy was deposited in one detector — amplitude error within 5 per cent across the calibrated range, saturated measurements flagged, dead time below 10 per cent at the maximum valid rate. But amplitude only means energy if something holds the peak still long enough to measure it. Rev A has comparators, a slow 16-bit converter for housekeeping, and no peak-hold or fast amplitude-capture path. The energy question has no hardware behind it yet, and nothing here should be read as spectroscopy.

The supporting electronics are drawn. The review was about the distance between that and an instrument whose counts would mean something.

The channels come first, and they have their own section below. Aperture, shielding and absorber geometry is not defined, so the channels have no basis for differing, and their responses are uncalibrated, so counts cannot become flux.

Then the measurement paths. The pulse-height architecture is incomplete. Coincidence timing needs verification against the 1 microsecond alignment requirement, and no hardware coincidence gate is drawn, so it rests on firmware. Threshold behaviour needs calibration across the DAC range.

Then the context. Magnetometer, IMU and temperature need synchronisation, and the magnetometer needs characterising against the payload's own currents. Spacecraft time, position and authoritative attitude have to come from the host. Storage roles for FRAM and flash are undefined, as is the integrity strategy.

Finally interference. Power converters, the RS-485 driver and the digital buses can all contaminate the measurement — REQ 8 requires that light, electrical noise, magnetic interference and crosstalk are not mistaken for particle events. Ground maps depend on calibrated response and quality flags that do not exist yet.

Four detector symbols are not four channels, so I went back to the sheet and checked.

The source is specific: each BPW34 shall have an independent bias, return path, amplifier, comparator and microcontroller input, with under 1 per cent crosstalk between channels.

Rev A draws four detector devices, but the schematic does not yet demonstrate four complete independent science channels from detector through amplification, discrimination and MCU input. The four BPW34 devices sit on a shared net rather than four separate bias and return paths. On the analog side there are three op-amp sections and three comparators, one of which drives a status LED rather than an event input. And the per-channel elements the source requires — feedback resistor, feedback capacitor, shaping network and calibration-injection point, four times over — are not there; the drawing carries four resistors and three capacitors in total.

So I am not going to call that four channels. What Rev A establishes is the shape of the instrument and most of its supporting subsystems. Completing the four independent paths is the first Rev B item, and every channel comparison depends on it.

A count with nothing attached to it is not a space-weather measurement. It is a number.

Turning it into one needs mission time, spacecraft position, attitude and therefore detector look direction, magnetic field, temperature, detector state, and live time with saturation and quality flags. Without position a rate cannot be placed on an orbit; without live time, a high rate is indistinguishable from a saturated instrument.

Rev A carries the local half: U20 magnetometer and U19 IMU at the required 10 Hz or better, U21 temperature, U14 for rails and bias, and U22 as backup timing. The authoritative half comes from the spacecraft — mission time within 10 milliseconds of the reference, and position and attitude within one second of the measurement for at least 99 per cent of accepted intervals.

The magnetometer needs a caveat. It measures the field where it sits, which includes whatever the payload's own converters and drivers produce. The requirement is residual error below 5 microtesla or the stricter spacecraft limit, verified by running the converters and interfaces while watching the magnetometer. It is not a clean measurement of the geomagnetic field until that has been characterised.

Counts become measurements through calibration, and this instrument needs more of it than most.

Seven quantities have to be established per channel: detector efficiency, geometric factor, energy response, angular response, the threshold actually in force, dead time, and the shielding response tying them together. Each is a term in converting a count rate into a flux, and none can be inferred from the schematic.

The methodology has two halves that check each other. Model the detector and shielding in a particle-transport program such as Geant4, which predicts the response for a geometry that has been defined. Then expose the instrument to calibrated electron or proton sources at an approved facility, across energies and incident angles. The criterion is agreement within approximately 20 per cent, or a larger measured uncertainty clearly reported. Then a blind exposure, reconstructed by the analysis team and checked against the facility's value — the test that catches a calibration agreeing with itself and nothing else.

No simulation has been run and no radiation exposure has taken place.

I started with the idea that more detectors would give me more information. Four seemed straightforwardly better than one.

The review took that apart. Detector count is not measurement diversity. Four detectors become four science channels only when each has a complete signal path, known geometry, controlled shielding, a calibrated response, synchronised timing, and channel identity preserved all the way to the ground. Drop any one and the extra detectors are extra copies — and I had drawn four devices without four paths.

Two other things changed. Shielding stopped being mechanical protection and became part of the measurement, because the aperture and absorber decide what each channel counts. And coincidence and pulse height turned out to answer different questions — whether a particle penetrated, and how much energy it left — which is not the same question twice, and neither identifies what the particle was.