CubeSTEM MissionLab Twin · M2-C

Three-Axis Laboratory Teacher Guide

Six software-only engineering laboratories built around fixed baseline-versus-intervention comparisons. The validated M2-A truth kernel remains unchanged and all sensor channels are simulated estimates.

Guided mode

Use the mission question and four sequenced steps. Learners explain evidence in plain language before changing level or seed.

Builder mode

Expose bounded controller and sensor configuration differences. Require learners to connect parameter changes to evidence channels.

Engineer mode

Require provenance, artifact hashes, criteria interpretation and an exported evidence report. This remains educational, not flight qualification.

Recommended 45–60 minute delivery pattern

Predict

Learners state the expected relationship before running the model.

Observe

Run the fixed baseline and record the relevant evidence channels.

Compare

Run one bounded intervention and inspect the criteria, plot and configuration difference.

Explain

Write a causal explanation and identify the fidelity boundary.

Laboratory 1

Attitude Acquisition: Tune for Faster Pointing

How do proportional gain, damping and torque limits change the time required to point a CubeSat?

25 min

Learning objective

Compare a deliberately weak baseline controller with a bounded tuned controller and explain the pointing-error trade-off.

Learner sequence

1.Predict whether stronger bounded control will reduce the final pointing error.
2.Run the fixed baseline and inspect its error, body-rate and wheel-speed evidence.
3.Run the candidate intervention and compare settling time, control effort and saturation.
4.Explain why a faster response is not automatically a safer or more efficient response.

Evidence channels

• pointing error
• body-rate magnitude
• wheel speed
• control effort
• settling time

Teacher prompt

Ask learners to distinguish response speed from stability, actuator stress and energy use.

Expected interpretation

A successful intervention should reduce pointing error without introducing avoidable wheel saturation.

Laboratory 2

Detumble Tuning: Remove Initial Body Rate

How does damping strength affect the removal of a high initial three-axis body rate?

25 min

Learning objective

Compare weak and stronger bounded damping while tracking final body rate, pointing error and wheel demand.

Learner sequence

1.Record the initial high-rate condition and predict the dominant axis.
2.Run the weak-damping baseline.
3.Run the candidate intervention and compare final body-rate magnitude.
4.Identify whether improved detumble performance creates higher wheel demand.

Evidence channels

• body-rate magnitude
• pointing error
• wheel speed
• control effort

Teacher prompt

Use the result to discuss why detumbling and precision pointing are related but different control objectives.

Expected interpretation

The candidate should remove more body rate while remaining inside the bounded wheel and torque model.

Laboratory 3

Reaction-Wheel Saturation: Diagnose the Limit

What evidence shows that an aggressive controller has pushed a reaction wheel to its modeled speed limit?

30 min

Learning objective

Compare a conservative baseline with an aggressive bounded candidate and identify saturation from multiple evidence channels.

Learner sequence

1.Run the conservative baseline and note peak wheel speed.
2.Run the aggressive candidate.
3.Locate saturation events and compare wheel speed, pointing error and applied torque.
4.Explain why more commanded torque cannot recover authority after the modeled wheel limit is reached.

Evidence channels

• wheel saturation events
• peak wheel speed
• applied torque
• pointing error

Teacher prompt

Require learners to cite at least two independent indicators before claiming saturation.

Expected interpretation

This is a diagnostic laboratory: the candidate is expected to reveal saturation, not to score as a universally better controller.

Laboratory 4

Sensor Noise and Bias: Separate Truth from Estimate

Can sensor quality change the estimated attitude while leaving the simulated spacecraft truth unchanged?

25 min

Learning objective

Hold the physics request constant while comparing ideal and degraded deterministic sensor profiles.

Learner sequence

1.Run the ideal-sensor baseline and record estimator error.
2.Run the degraded candidate using the same physics seed and controller.
3.Confirm that the truth artifact remains unchanged.
4.Explain why estimated telemetry must never be labelled as measured truth.

Evidence channels

• truth artifact hash
• mean estimator error
• maximum estimator error
• provenance

Teacher prompt

Ask learners to identify which channels can change without changing the underlying spacecraft motion.

Expected interpretation

The truth hash should remain identical while estimator error increases under the degraded sensor profile.

Laboratory 5

Estimator Dropout: Observe Gyro-Propagation Fallback

What happens when one or both reference-vector measurements are unavailable?

30 min

Learning objective

Compare classroom and challenging sensor profiles and quantify degraded estimator frames.

Learner sequence

1.Run the classroom baseline and count degraded estimator frames.
2.Run the challenging candidate with deterministic vector dropouts.
3.Compare maximum estimator error and fallback count.
4.Explain why fallback continuity is useful but does not make the estimator flight certified.

Evidence channels

• degraded estimator frames
• maximum estimator error
• sensor availability
• provenance lineage

Teacher prompt

Discuss observability and why a propagated attitude can drift when absolute references disappear.

Expected interpretation

The challenging profile should expose more fallback frames and usually higher estimator error.

Laboratory 6

Disturbance Rejection: Hold Pointing Under Torque

How does controller tuning affect pointing performance under a constant modeled disturbance torque?

30 min

Learning objective

Compare weak and tuned bounded control under the same constant disturbance.

Learner sequence

1.Run the weak-control baseline under the fixed disturbance.
2.Run the tuned candidate with the same initial condition and seed.
3.Compare mean and final pointing error, wheel demand and control effort.
4.State what additional orbit, magnetic and environmental models would be required before making a flight claim.

Evidence channels

• mean pointing error
• final pointing error
• wheel speed
• control effort

Teacher prompt

Use the disclosure boundary to distinguish a controlled teaching disturbance from an orbit-derived environment.

Expected interpretation

The tuned candidate should reject more of the fixed teaching disturbance while remaining bounded.

Evidence rubric · 12 marks

Prediction

States a testable expected relationship before running the laboratory.

Evidence selection

Uses at least two relevant channels and cites baseline and candidate values.

Engineering explanation

Connects the bounded intervention to the observed response using correct terminology.

Provenance

Correctly distinguishes simulated truth, estimated sensors, commanded control and derived metrics.

Fidelity boundary

States what the model does not prove and avoids flight or hardware claims.

Mandatory fidelity statement

M2-C uses the validated deterministic F2 attitude kernel, controlled teaching vectors, fixed modeled disturbances and deterministic sensor estimates. It does not provide orbit-derived environment, RF, integrated EPS, thermal, payload, hardware-command or flight-qualification authority.

Measured hardware channels: 0
Official attempts: disabled
Mission Credits: not used
Arbitrary learner code: prohibited