Lecture 02 02 Sensitivity

Main ideas / concepts covered (measurement error & uncertainty)

The lecture explains that measurement accuracy depends not only on the measuring instrument, but on the measurement system (instrument + setup + method + conditions). It then introduces several instrument/system performance terms that are often confused or misused.

---

Key concepts and definitions

1) Accuracy vs. Precision (and common misuse)

• Accuracy: how close readings are to the true/actual value.
• Precision: how consistently repeated readings match each other.

The speaker notes these ideas are frequently mixed up with instrument-related and system-related properties.

---

2) Repeatability (often “repeatability/ability to repeat”)

Focus: same method, same conditions, short time, repeated measurements.

Repeatability is the ability of a measurement system/instrument to measure the same input repeatedly within a short time.

Example (boiling water):

• Boiling water is ~100°C and kept at that temperature continuously.
• You measure temperature 10 times within a 30-second interval.
• If readings stay near 100.0, 100.1, 99.9, etc., the system is repeatable.
• If readings drift noticeably during the short measurement window (e.g., 100.0 → 100.2), repeatability is limited.

---

3) Reproducibility (often “reproducibility/ability to reproduce”)

Focus: different time/people/method, but same intended measurement result.

Reproducibility is the ability of the measurement system to produce essentially the same results when:

• the measurement is done at a different time
• potentially by different people
• potentially with a different measurement procedure/method

Example (physics lab groups):

• Group A measures a length 5 times and gets roughly the same set of results (e.g., around 100–110 with small variation).
• Next week Group D repeats the measurement using the same general setup/instrument.
• If Group D gets a similar result range, that demonstrates reproducibility.

---

4) Tolerance / Maximum error (“maximum permissible error”)

Tolerance is related to accuracy:

• Tolerance (Maximum error): the maximum allowed deviation between:
  • the reading value and
  • the true value.

Example (thermometer / human temperature):

• If a thermometer indicates values like 37.5°C, 37.0°C, 37.7°C around a nominal expected body temperature,
• the difference from the expected/true value is treated as the error (within the device’s tolerance).

---

5) Measurement range (span / operating limits)

Focus: maximum and minimum values the instrument can measure.

It is the instrument’s allowed/usable measurement interval.

Examples:

• Body thermometer: might measure only within a certain interval (e.g., 25°C to 45°C).
• Ruler: typical everyday rulers have a marked range (e.g., 0–30 cm).

---

6) Linearity

Focus: how well the relationship between actual value and measured value follows a straight (linear) trend.

Linearity means that when plotting:

• actual input (x-axis) vs measured output/reading (y-axis),

the points fall near a straight line.

Example (warming water to boiling):

• Actual values increase from room temperature to boiling (e.g., 30°C → 50°C → 60°C → 100°C).
• Instrument readings increase correspondingly (e.g., 30.5, 52, 59.05, 100.4, etc.).
• A linear regression yields a “best-fit” line.
• Linearity error is described as the distance between the measured point and the actual (true) value or expected line behavior.

---

7) Overall accuracy (decomposed)

The lecture claims overall instrument accuracy has additional components, particularly:

• Sensitivity
• Consistency (related to repeatability)

---

8) Sensitivity

Focus: how much the instrument reading changes when the input changes.

Sensitivity is defined via the response to environmental/condition changes.

Example (body temperature / sensors):

• If an infrared thermometer shows only one digit after the decimal, it may not detect very small changes well.
• Rule-of-thumb mentioned:
  • for high-detail industrial applications, sensitivity may require more digits (e.g., two digits behind the decimal place) to resolve changes like 0.05°C.

---

9) Consistency (measurement “uniformity” across repeated trials)

Focus: repeated measurements of the same object under the same conditions yield the same output.

This aligns with repeatability: stable output across repeated measurements.

---

10) Hysteresis

Focus: difference in readings depending on measurement direction (rising vs falling).

Hysteresis is the difference between:

• measurements during heating (e.g., 30°C → 100°C) and
• measurements during cooling (e.g., 100°C → 30°C)

The lecture explains that:

• the temperature curve shape can differ between heating and cooling,
• and that difference affects measurement results.

This effect is described as a characteristic of the measuring instrument.

---

Methodology / “instruction-like” guidance mentioned (embedded in examples)

No explicit step-by-step procedure for performing uncertainty calculations is provided. However, the lecture implies an evaluation approach for instrument qualities:

• To assess repeatability:
  • Keep input, location, method, and conditions the same.
  • Perform multiple readings within a short time.
  • Check whether results remain close across trials.
• To assess reproducibility:
  • Repeat the same measurement at a different time (often by different people / under slightly different execution).
  • Verify whether results remain similar to prior trials.
• To assess range:
  • Confirm the instrument’s stated minimum and maximum measurable values.
• To assess linearity:
  • Measure known actual inputs across the operating region.
  • Fit a linear regression line and evaluate deviations/error.
• To assess sensitivity:
  • Apply small known changes in input/environment.
  • Determine whether the output changes proportionally and with sufficient resolution.
• To assess hysteresis:
  • Measure while ramping input upward, then ramp downward.
  • Compare readings and note differences.

---

Speakers / sources featured

• One lecturer/speaker (no name provided in the subtitles).
• No other identifiable external sources or speakers mentioned.