Precision propulsion metrology
Design force and momentum experiments that can survive thermal, electromagnetic, vibration, gas, cable, and analysis artifacts.
Develop torsion balances, pendula, interferometers, accelerometers, calibration transfer, environmental telemetry, impulse and momentum accounting, blinded reversals, artifact injection, multi-lab replication, and decision thresholds for extraordinary propulsion claims.
Before you begin
- • Experimental methods and error analysis
- • Electromagnetism and mechanics
- • Signal processing and statistical inference
By the end, you can
- • Build a traceable force and impulse calibration chain.
- • Forward-model thermal, magnetic, electrostatic, vibration, gas, and cable artifacts.
- • Use blinded modulation and reversal to separate mechanisms.
- • Coordinate an independent multi-lab replication with shared standards.
Interactive model
Explore before calculating
Live laboratory
Blinded propulsion round-robin coordinator
Compare three independent force reports with a hidden calibration injection, shared uncertainty contract, heterogeneity test, and sham residual. Agreement, calibration recovery, and artifact rejection must pass together.
Weighted mean: 50.33 nN
Mean uncertainty: 2.89 nN
Recovery residual: 0.12σ
χ²/dof: 0.17
Largest pair spread: 0.57σpair
Sham residual: 0.60σ
Calibration round robin passes the declared recovery, heterogeneity, and sham gates.
This equal-uncertainty calibration exercise is not a propulsion-effect meta-analysis. A claim run also needs blinded active/sham schedules, correlated covariance, transfer functions, momentum closure, all-run reporting, independent apparatus, preregistered scaling, and a random-effects model when between-lab variance is nonzero.
Level 6 · Research preparation teaching kit
Record the investigation. Teach the reasoning.
A learner-facing lab record and a course-specific instructor guide turn the live model into a repeatable classroom investigation.
Learner record
Blinded propulsion round-robin adjudication
Can multiple laboratories recover calibrated injections, reject shams, and agree within declared heterogeneity before interpreting an unknown force residual?
Download learner recordInstructor guide
Teach for evidence, not button pushing
Researchers validate calibration, artifact rejection, blinding, and cross-laboratory agreement before attributing precision force residuals.
Download instructor guideAdvanced assessment
Reconstruct it. Quantify it. Try to break it.
Treat calibration transfer, laboratory heterogeneity, injections, and shams as prerequisites to any propulsion interpretation. Three research-level challenges include explicit deliverables and scoring criteria.
Portable research dataset
Record data that another laboratory can open.
Blinded laboratory calibration, injection, sham, residual, and qualification records. JSON preserves schema and provenance; CSV supports ordinary analysis tools. Imports stay in this browser and are limited to 1 MB and 5,000 records.
Ready for a new research record.
| Laboratoryidentifier | Blind statelabel | ResidualN | UncertaintyN | Injection recoveredboolean | Sham rejectedboolean | Qualifiedboolean | Record |
|---|---|---|---|---|---|---|---|
Schema field definitions
- Laboratory · identifier
- Stable anonymized laboratory ID.
- Blind state · label
- Unknown, injection, or sham after authorized unblinding.
- Residual · N
- Calibrated force residual.
- Uncertainty · N
- Combined declared uncertainty.
- Injection recovered · boolean
- Validation result.
- Sham rejected · boolean
- Validation result.
- Qualified · boolean
- Prespecified laboratory qualification decision.
Lesson 1 of 3
Force sensors, transfer functions, and calibration
How does a raw displacement become a force with traceable uncertainty?
Torsion fibers, pendula, flexures, and interferometers have frequency-dependent transfer functions. Calibration must match signal bandwidth, geometry, amplitude, and operating conditions.
Electrostatic, gravitational, radiation-pressure, or mechanical standards provide known inputs. Drift and nonlinearity require repeated bracketing calibrations rather than one conversion factor.
Worked example
A balance has stiffness 2×10⁻⁵ N/rad and deflects 3 μrad. Estimate force-equivalent torque for unit lever arm.
- 1. Multiply stiffness by angle.
- 2. Use consistent radians.
- 3. For a 1 m lever arm, torque and force share the numerical value.
6×10⁻¹¹ N for the stated idealized geometry.
Try it
Calibration transfer audit
Materials: Synthetic calibration and signal sweeps
- 1. Fit transfer function.
- 2. Check linearity and hysteresis.
- 3. Interpolate at signal frequency.
- 4. Propagate calibration covariance.
Notice: A calibration far from the signal's frequency or geometry can create a false residual.
Check your understanding: Why bracket experimental runs with calibrations?
Answer: To measure drift and verify sensitivity before and after the claimed signal.
A pre-run calibration alone cannot rule out later transfer-function change.
Lesson 2 of 3
Momentum paths and artifact injection
Which ordinary mechanism can produce the same sign, timing, and scaling as the target?
Thermal expansion, radiometric forces, outgassing, convection, cable forces, electromagnetic coupling, vibration rectification, center-of-mass shifts, and data-window choices can mimic thrust.
Artifact injection deliberately exaggerates each mechanism to measure its transfer function. Reversal and sham conditions should modulate the proposed cause while holding artifacts or vice versa.
Worked example
A force reverses with device orientation but the power cable also bends oppositely. What control is decisive?
- 1. Route a symmetric or wireless power path.
- 2. Use a matched heater sham.
- 3. Measure cable force independently.
- 4. Blind orientation labels during analysis.
Orientation reversal alone is not mechanism-specific when the cable geometry reverses too.
Try it
Artifact challenge campaign
Materials: Apparatus digital twin and injected nuisance signals
- 1. Create one injection per artifact.
- 2. Recover transfer functions blindly.
- 3. Set sensor and shielding requirements.
- 4. Test whether the primary estimator rejects them.
Notice: A pipeline validated only on desired signals may be optimized to miss real artifacts.
Check your understanding: What does a null result during a matched heater test constrain?
Answer: Thermal artifacts only to the extent that the heater reproduces the device's relevant temperature and mechanical field.
A poor thermal surrogate gives a weak bound.
Lesson 3 of 3
Blind analysis, momentum closure, and multi-lab replication
What protocol would make a positive residual credible outside the originating laboratory?
A preregistered primary estimator, hidden active/sham schedule, frozen exclusions, and injected canaries reduce analysis flexibility. All acquired runs, including failed and null runs, belong in the package.
Independent replication changes investigators and apparatus while preserving the predicted modulation. A momentum ledger and environmental telemetry must close within uncertainty across facilities.
Worked example
Three labs measure 2.0±0.8, 0.1±0.7, and −0.2±0.9 μN. Is replication established?
- 1. Check individual consistency with zero.
- 2. Combine with inverse-variance weighting.
- 3. Test heterogeneity.
- 4. Compare with preregistered replication threshold.
One suggestive result with two nulls does not establish replicated thrust; the combined evidence and heterogeneity must be reported.
Try it
Round-robin protocol
Materials: Shared calibration artifacts and blinded device set
- 1. Freeze protocol centrally.
- 2. Distribute blinded standards/devices.
- 3. Run local analyses before unblinding.
- 4. Pool all data with a declared meta-analysis.
Notice: Shared standards reveal calibration differences while independent apparatus tests laboratory-specific artifacts.
Check your understanding: Why publish failed and null runs?
Answer: To expose selection effects and estimate the true false-positive and variability structure.
Selective reporting inflates apparent reproducibility.
Formula-to-meaning deck
Read the equation in ordinary language.
x(ω)=H_F(ω)F(ω)+ΣH_i(ω)n_i(ω)
Measured displacement combines force response with frequency-dependent nuisance couplings.
I=∫F(t)dt=Δp
Impulse must correspond to momentum transferred somewhere in the complete system.
μ̂=Σw_i y_i/Σw_i, w_i=1/(σ_i²+τ²)
A random-effects synthesis combines lab results while allowing between-lab variation.
Independent practice
Problem set
Work each problem before opening its hint and solution.
1. A 2 μN force lasts 0.5 s. What impulse is claimed?
Reveal hint
Integrate constant force over time.
Reveal solution
1 μN·s, or 10⁻⁶ N·s.
2. A cable artifact transfer is 0.3 μN/mm and motion is 0.02 mm. Estimate artifact.
Reveal hint
Multiply.
Reveal solution
0.006 μN.
3. Why is a 5σ single-lab result not sufficient for extraordinary propulsion?
Reveal hint
Consider systematic uncertainty and independence.
Reveal solution
Statistical significance does not cover unmodeled artifacts, analysis flexibility, or laboratory-specific systematics; independent controlled replication is required.
Derivation studio
Build the result, line by line.
Keep the assumptions visible so the mathematics remains auditable.
Starting point
Torsion-balance transfer function
Iθ̈+bθ̇+κθ=τ(t)
- 1. Fourier transform time derivatives.
- 2. Collect terms multiplying θ(ω).
- 3. Solve for angular response.
- 4. Relate torque to force by lever arm.
H_F(ω)=L/[κ−Iω²+i bω] for the stated geometry
Signal amplitude and phase depend on frequency, damping, inertia, stiffness, and lever arm.
Starting point
Correlated uncertainty propagation
Force estimate F=f(x,c_1,…,c_n)
- 1. Linearize with the Jacobian.
- 2. Construct the full covariance matrix.
- 3. Include shared calibration and drift correlations.
- 4. Compute JΣJᵀ.
σ_F²=JΣJᵀ
Treating correlated calibration terms as independent can dramatically overstate significance.
Computational notebook
Turn the model into an experiment.
Precision-force digital twin
Can the preregistered estimator recover target impulses while rejecting realistic artifact families across laboratories?
Inputs
- • Mechanical transfer function
- • Environmental channels and injected artifacts
- • Blind active/sham schedule
- • Calibration covariance
Algorithm
- 1. Simulate apparatus and nuisances.
- 2. Fit transfer functions on training injections.
- 3. Freeze primary estimator.
- 4. Evaluate blinded detection, false alarms, and momentum closure.
Evidence to produce
- • Transfer and artifact models
- • Blind recovery scorecard
- • Cross-lab uncertainty and momentum ledger
Paper-reading studio
Interrogate the source, not its reputation.
Reconstruct the assumptions, reproduce one calculation, and stop at the boundary of the reported evidence.
Extraordinary propulsion result audit
Which measured residual remains after reproducing calibration, artifact transfer, momentum, and selection accounting?
- 1. Reconstruct apparatus dynamics.
- 2. Recompute calibration and covariance.
- 3. Forward-model thermal/electromagnetic artifacts.
- 4. Include every run and replication.
Calculation to reproduce: Reproduce the primary force estimate, complete uncertainty, and one dominant artifact bound.
Evidence boundary: A residual force is not automatically anomalous propulsion; mechanism identification requires momentum closure, discriminating scaling, and independent replication.
Graduate oral defense
Defend a bounded claim under pressure.
Argue the strongest support, state the strongest objection fairly, and identify evidence that could actually decide the issue.
Proposition
Precision propulsion metrology should prioritize falsification capacity over maximum claimed sensitivity.
- 1. Artifact injections reveal real nuisance transfer functions.
- 2. Reversals and blinded shams test causal specificity.
- 3. Independent labs bound hidden apparatus dependencies.
Strongest objection: Aggressive controls can reduce sensitivity or eliminate unconventional configurations before their signal is understood.
Deciding evidence: A signal that survives prespecified artifact challenges and replicates with predicted scaling across materially different apparatus.
Research practicum
Make the work inspectable before making it impressive.
Pre-register the decisive test, package every dependency, and pass explicit milestone gates before interpretation expands.
Coordinate a blinded propulsion round robin
Does the target signal replicate across independent instruments with closed momentum and artifact ledgers?
Preregister
- • Freeze modulation, estimator, exclusions, and threshold.
- • Define shared calibration and artifact injections.
- • Specify meta-analysis and heterogeneity handling.
Reproducibility package
- • Apparatus drawings and transfer functions
- • Raw synchronized telemetry
- • Calibration and injection files
- • Immutable analysis environment and all-run manifest
Milestone gates
- 1. Single-lab canary validation
- 2. Blinded internal replication
- 3. Independent apparatus replication
- 4. Joint unblinding and public report
Continue into the evidence