Rulebook — Numerical Clock Networks

This page contains the rules, assessment criteria, and procedures that apply to this experiment. For the full, programme-wide FP rulebook, see the master document.

Laboratory Steps

Lab hours: 9:00–17:30 h. Coordinate details with your tutors.

Step	Format	What you demonstrate
1 Entrance Session	15-min presentation + 30-min discussion	Key physics of clocks, noise types, comparison geometry; tutorial solutions
2 Active Lab + Lab Notes	Throughout experiment	Engaged simulation work; in-lab analysis; complete lab notes
3 Findings Session	~30-min presentation + 15-min discussion	Key findings with uncertainties; comparison to expectations; unresolved questions

Limit: More than 2 failed steps (across all experiments in the course) results in course failure.

Step 1: Entrance Session

Present the key physics: what is a clock, what are the noise types, what does η(τ) control? Demonstrate that you have completed the Tutorial and can explain your solutions.

Outcomes: Passed (proceed) · Repeated (gaps clarified, proceed, documented) · Failed (reschedule, counts toward limit).

Step 2: Active Lab + Lab Notes

Lab Conduct

Stop and ask if uncertain. Articulate actual reasoning, not what you think is expected.
Push notebooks to your repository fork regularly.

In-Lab Analysis — Essential

Estimate results as you go. Quick Allan deviation plots reveal whether your simulation works.
Check noise models early. If the σ_y(τ) slope is wrong, debug now — not during report writing.
Compare to expectations. Orders-of-magnitude discrepancies require investigation before proceeding.
Plot data during simulation. Trends, outliers, and convergence failures become visible immediately.

Key principle: Problems caught during simulation can be fixed. Problems discovered during report writing cannot.

Lab Notes Requirements

Your lab notes must allow verification and reconstruction of your work. Record: simulation parameters and seeds, code versions (git commits), all intermediate results, anomalies as they occur, reasoning for parameter choices.

Tutor Guidance

Default: Observe. Students lead.
Guide when stuck: Clarifying questions and directions — not solutions.
Intervene only under: T1 (Safety), T2 (Clear factual error, >90% confidence), T3 (Process feedback).

Step 3: Findings Session

Present key findings with uncertainties, comparison to expectations, and unresolved questions (~30 min + 15 min discussion).

Minimum requirement: At least a qualitative discussion of findings with uncertainties, comparison to expectations, and unresolved questions. Quantitative treatment strengthens the presentation.

Assessment: Seminar Experiment Model

This experiment follows the seminar-experiment assessment model.

Component	Weight	Deadline
Short Report	30%	7 calendar days before seminar
Seminar Presentation	70%	Scheduled date
Weight in lab grade	10–30% of the overall Master Laboratory grade, depending on time effort. Agreed individually with the organiser before the learning phase begins.

Short Report Requirements

The short report documents the complete analysis pathway. Bullet points and keywords are acceptable; the seminar provides the venue for full verbal articulation.

Criterion	Requirement
Data Analysis	Complete pathway documented; all steps from noise generation to final η_opt map traceable
Uncertainty Treatment	Uncertainty propagation documented; frequentist CIs and Bayesian credible intervals; sources identified
Assumptions	All assumptions and approximations explicitly listed (noise models, prior choices, link models)
Comparison to Expectations	Analytic predictions stated; deviations noted and addressed
Key Results	Final numerical results clearly marked with uncertainties
Attachments	Lab notes, simulation code (git link), raw output data

Late Submission

Submission	Consequence
On time	Seminar proceeds as scheduled
1–3 days late	0.3 grade penalty; seminar may be rescheduled
>3 days late	0.7 grade penalty; seminar rescheduled
Not submitted by rescheduled date	Combined grade 5.0

Extensions: Written request before deadline with documentation (illness, family emergency, comparable). Workload alone does not constitute grounds.

Seminar Presentation (60 minutes)

Phase	Duration
Uninterrupted presentation	~30 min
Open discussion	~15 min
Exam questions (organisers/tutors)	~15 min

Content Requirements

Motivation: What is a clock? Why comparison geometry?
Theory: Frequentist and Bayesian foundations; the η(τ) parameter.
Methods: Simulation setup, noise models, network topologies, software tools.
Data analysis: From raw time series to Allan deviation to η_opt map.
Results: Final η_opt landscape with uncertainties; model comparison outcomes; closure diagnostics.
Summary: What worked, what broke, what you would do differently.
Sources: Cited on slides.

Presentation Quality

Coherent logical thread from question to result.
Claims follow from evidence. Uncertainties acknowledged.
Results traceable to submitted short report.
Consistent labels and symbols; avoid unexplained abbreviations.
Address peers, not just examiners.

Participation Requirement

Enrolment: Only students properly enrolled in the M.Sc. Applied Physics programme (or equivalent, confirmed by the examination office) may participate. Enrolment status must be verified before the first session.
Workspace conduct: Shared office and computing resources must be treated with care. Leave workstations as you found them.

Numerical lab: This experiment is entirely office-based and computational. No laser, radiation, or laboratory safety training is required. Standard university workspace rules apply.

Escalation

Attempt resolution with tutor (unless the tutor is the subject of the concern).
If unresolved within 7 days: email organiser.
If organiser unresponsive: contact Dean of Studies.

Response deadlines: Acknowledgement within 2 working days; resolution within 10 working days. Good-faith escalation cannot lead to disadvantage.

Consequences

Situation	Consequence
Participation without valid enrolment	Immediate exclusion; results void
>2 failed steps (course-wide)	Course failure
Late short report ≤3 / >3 days	0.3 / 0.7 grade penalty
Short report + seminar combined 5.0	Resubmission + repeat seminar opportunity

Core Commitments

Students: Prepare thoroughly. Present honestly. Ask when uncertain. Meet deadlines or communicate early.

Tutors: Observe by default. Guide when stuck. Intervene only for safety, clear errors, or process feedback. Assess against published criteria only.

The system: All criteria are known before you begin. Every decision has a name and reason. Mistakes are correctable. Divergence is legitimate.

Appendix: Lab Notes Checklist

Title and date for each session.
Objective: state the purpose of the day’s work.
Simulation parameters: noise amplitudes, number of samples, random seeds, git commit hash.
Procedure: what you ran, in what order, including changes from the plan.
Observations and data: all output, including intermediate plots and unexpected behaviour.
Calculations and analysis: Allan deviations, posterior summaries, closure residuals with uncertainties.
Interpretation: what the results mean in terms of the framework.
Errors and anomalies: convergence failures, numerical artefacts, unresolved discrepancies.
Improvements: what you would change next time.
References: cited where used.
Sign and date each entry.

Format: You choose, but a single file uploaded to ILIAS. Jupyter notebooks with embedded markdown are encouraged — they combine code, figures, and narrative in one traceable document.