Temporal Instability: What to Log and Why

  • Technical Note / Measurement Methods
  • Title: Irradiance Non-Uniformity: Measurement Setup Tips
  • 1-line summary: Common pitfalls and a simple checklist for spatial uniformity measurements.
  • Date: 2026-02-22
  • Spectrum: General
  • Level: Standard
  • Category: Measurement Methods
  • Related: Temporal Instability: What to Log and Why / (Example) Non-Uniformity Report Template

1. Purpose

Many users say “the simulator is stable,” but the conclusion depends on what was logged and how it was logged.
This note provides a practical logging set so that:

  • short-term fluctuation (seconds–minutes) and longer drift (hours) are not mixed,
  • causes (temperature, warm-up, adjustments) can be separated from true source drift,
  • results can be compared across days, instruments, and sites.

2. Definitions (quick)

2.1 Temporal instability (of irradiance)

A time-dependent change in irradiance during operation. It includes:

  • Noise / fluctuation: rapid variation (seconds–minutes)
  • Warm-up transient: systematic change after turn-on or shutter-open
  • Drift: slow monotonic change over longer time (tens of minutes–hours)
  • Step / event: sudden change caused by a discrete event (adjustment, shutter, lamp re-ignite, sensor reposition)

2.2 Stability vs drift (practical separation)

  • Stability = “How much it varies around a level” within a defined time window.
  • Drift = “How the level shifts” over a longer window.

Key point: You cannot claim stability/drift without stating (a) time window, (b) sampling interval, (c) what events occurred.

3. Common misunderstanding (and the fix)

3.1 Typical misunderstandings

  • Logging only irradiance, then attributing any change to “lamp drift.”
  • Not recording warm-up start/end, then comparing early data with steady-state data.
  • Making adjustments (e.g., 1 SUN re-tune) but not logging the adjustment time → drift plot becomes meaningless.
  • Ignoring detector and ambient temperature, even though sensitivity can be temperature dependent.
  • Using “average over 1 hour” without reporting the raw sampling interval or missing spikes.

3.2 Simple reporting rule (always state these 6 items)

Whenever you report temporal instability/drift, state:

  1. Time window: ______ (e.g., 0–60 min / 2–24 h)
  2. Sampling interval: ______ (e.g., 1 s / 10 s / 1 min)
  3. Warm-up definition: ______ (e.g., shutter-open time as t=0; warm-up = first 20 min excluded)
  4. Events/operations: ______ (adjustments, shutter close/open, re-ignite, sensor move)
  5. Measurement basis: instrument / range / settings / traceability: ______
  6. Temperature logs: sensor temp and ambient (or “not logged”): ______

4. What to log (recommended logging items)

Use the following three tiers.

4.1 Minimum (must-have)

Common misunderstanding (and the fix)

  • Irradiance value (with unit and whether it is total / band-limited)
  • Timestamp (absolute time recommended)
  • Sampling interval (fixed, documented)
  • Start reference event (e.g., “shutter opened” as t=0)

4.2 Recommended (for meaningful diagnosis)

  • Detector temperature (or sensor body temperature if available)
  • Ambient temperature near DUT plane
  • Source operation state (lamp power mode, current setpoint, shutter status)
  • Event markers: adjustment time, re-tune, re-ignite, interlock trip, etc.
  • Notes field (short text): “operator action / reason”

4.3 Advanced (if you want root-cause capability)

  • Power supply key parameters (current, voltage, regulation mode, alarms)
  • Optical head temperature (filter housing, lamp house if available)
  • Airflow / fan state (if it changes automatically)
  • Reference channel (secondary sensor for cross-check)

4.4 Practical tip

If you can only add one thing beyond irradiance: add event markers + detector temperature.
Without events, you cannot interpret steps. Without temperature, you cannot separate sensitivity effects from true irradiance drift.

5. Quick checklist table

ItemMinimumRecommendedWhy it matters
Irradiance value✓✓Primary performance metric
Timestamp (absolute)✓✓Aligns with events and comparisons
Sampling interval✓✓Defines what “stability” means
Start event (t=0)✓✓Avoids warm-up mixing
Event markers (adjust/re-ignite/shutter)✓Explains steps / discontinuities
Detector temperature✓Separates sensor drift vs source drift
Ambient temperature✓Helps explain slow trends
Source state (mode/setpoint)✓Makes repeatability defensible
PSU parameters / alarmsRoot-cause analysis

6. Notes (SAN-EI recommendation)

  • Always present a drift plot together with the event timeline (even if it is just a list of timestamps).
  • If adjustments are performed, report results as two segments: “before adjustment” and “after adjustment.”
  • If a reader might misinterpret, include a one-sentence disclaimer:
    “Observed steps coincide with operator adjustments; they are not intrinsic drift.”

Return to INDEX

MENU