6.3 IR Window Transmittance Corrections
Key Takeaways
- IR inspection windows (polymer/polycarbonate and crystal materials) transmit only a fraction of infrared; that fraction is the window transmittance τ and must be entered on the camera
- Correct method: set the camera’s external optics / window transmittance parameter so the radiometric inversion accounts for τ — do not subtract a fixed number of degrees from the displayed temperature by hand
- An 80% transmittance window (τ = 0.80) means about 20% of the in-band radiance is lost at the window (plus reflection/emission effects the camera model may include); leaving τ = 1.00 under-reads hot targets
- Soiling, scratches, condensation, and non-normal viewing angles reduce effective transmittance below the clean-window specification
- Window correction is in addition to the five free-path parameters; it does not replace emissivity or RAT
Energized electrical equipment is often inspected through infrared (IR) inspection windows — sealed viewports that transmit infrared while maintaining enclosure integrity and supporting safer standoff. Level II thermographers must treat the window as an optical element in the radiometric chain, not as invisible glass. The controlling parameter is window transmittance (τ).
Why Windows Exist
| Benefit | Explanation |
|---|---|
| Reduced open-door exposure | Inspect without fully opening energized gear when procedures allow window use |
| Arc-flash / PPE policy alignment | Supports distance and barrier strategies defined by the facility |
| Repeatable aiming points | Fixed ports standardize routes and trending geometry |
| Environmental sealing | Limits dust and accidental contact with live parts |
Windows do not remove the need for qualified electrical safety practices. They also do not transmit 100% of IR.
Window Materials: Polymer vs Crystal
IR windows used in electrical gear fall into broad families:
| Material family | Typical traits | Transmittance notes |
|---|---|---|
| Polymer / polycarbonate IR optics | Impact resistant, common in industrial ports | τ often specified in a band (e.g., LWIR); may be ~0.5–0.9 depending on design and coatings — use the manufacturer value for that model and band |
| Crystal (e.g., specialized IR crystals used in some ports) | Optical performance designed for IR bands | Often higher or more stable τ in specified band; still not 1.00; mechanical properties differ |
| Ordinary building glass / viewing glass | Not an IR window | Typically opaque or very poor in LWIR — do not treat as an IR port |
Always use the manufacturer’s transmittance specification for your camera’s spectral range. A number printed for MWIR may not apply to LWIR. If the window is labeled τ = 0.80 at 8–14 µm, that is the starting clean-window entry for a matching LWIR imager — until dirt or damage changes reality.
Transmittance vs other optical losses
A real window:
- Transmits fraction τ of incident IR (idealized single-pass description)
- Reflects some energy from each surface
- Absorbs some energy and can emit according to its own temperature
Camera “external optics” or “window” modes implement a correction model. Your job is to supply the correct τ (and follow the vendor’s prompts for window temperature if asked). Do not invent a personal shortcut that ignores the model.
The Golden Rule: Enter τ on the Camera — Do Not Arithmetic-Correct Temperature
Correct: Set the camera parameter for window/external optics transmittance to the window’s value (for example 0.80), complete the five free-path parameters, then read T_obj from the radiometric image.
Incorrect: Read an uncorrected temperature of 80 °C and “fix” an 80% window by computing 80 / 0.80 = 100 °C on a calculator, or by adding a fixed +5 °C “window fudge factor.”
Why hand arithmetic on temperature fails
Radiance and temperature are linked nonlinearly (Planck / calibration curves; recall T⁴ intuition from Chapter 4). Cutting radiance by 20% does not cut temperature by 20%. Also, the detector still receives a mix of:
- Attenuated target emission
- Attenuated reflected scene
- Window emission/reflection
- Atmosphere on both sides of the port as applicable
Only the camera’s radiometric inversion, fed with τ, applies a consistent physical correction across the temperature range. Post-hoc degree adjustments break that consistency and fail exams and audits.
| Approach | Acceptable? | Reason |
|---|---|---|
| Enter τ = 0.80 in camera window parameter | Yes | Model-based inversion |
| Software batch reprocessing with same τ if supported | Yes, if radiometric | Equivalent to correct parameter entry |
| T_corrected = T_reading / τ | No | Treats temperature as if it scaled like radiance |
| T_corrected = T_reading + 10 °C always | No | Fixed offset ignores physics and target temperature |
| Raise ε to “soak up” window loss | No | Wrong physical parameter; corrupts reflection logic |
The 80% Window Example
Suppose a polymer IR window is specified at τ = 0.80 (80% transmittance) in your LWIR band when clean and viewed near normal incidence.
Setup
| Parameter | Example value |
|---|---|
| Window transmittance | 0.80 |
| Emissivity (tape on terminal) | 0.95 |
| RAT | 25 °C |
| T_atm | 25 °C |
| RH | 40% |
| Distance | 2.0 m (to target plane per procedure) |
What τ = 0.80 means
Approximately 80% of the relevant IR from the scene side is transmitted through a clean window at the design conditions; about 20% is not transmitted (reflected/absorbed per the optic). The camera must be told τ = 0.80 so it does not assume a free view.
If you leave τ = 1.00 (wrong)
The imager assumes no window loss. For a hot connection, less radiance arrives than a free-path hot target would send, so the uncorrected solution typically reports a temperature that is too low. That can under-rank severity (for example, missing a P2/P1 threshold) — a serious reliability and safety miss.
If you enter τ = 0.80 (correct)
The inversion attributes the reduced radiance partly to the window and solves for a higher object temperature consistent with the model — closer to true surface temperature (assuming ε, RAT, atmosphere, and spot size are also correct).
Numeric discipline on exams
Questions may say: “Window transmittance is 80%. What should the thermographer do?”
Best answer pattern: Enter 0.80 (80%) as the IR window / external optics transmittance on the camera (and keep proper ε and RAT). Not: multiply or divide the displayed °C by 0.8.
If options include both “set transmittance to 0.8” and “divide temperature by 0.8,” choose the parameter entry.
Soiling, Damage, and Condensation
Published τ assumes a clean, undamaged window used as specified.
| Degradation | Effect on effective τ | Field action |
|---|---|---|
| Dust, grease, fingerprints | Lowers τ; may add scatter | Clean per manufacturer; re-inspect |
| Scratches / pitting | Scattering, lower effective τ | Replace window if optical quality lost |
| Condensation / films | Strong IR absorption possible | Dry/clear; do not force quantitative claims through fogged ports |
| Paint overspray / stickers | Blocks IR | Remove obstruction; never measure “through” opaque coverings |
| Unknown after years of service | τ no longer matches label | Verify with known reference if critical; maintain PM on windows |
Exam idea: A dirty window is not fixed by raising emissivity. Clean or replace; if you must work, understand that effective transmittance is lower than the nameplate — quantitative risk rises.
Viewing Angle
IR windows and target emissivity both care about angle.
| Geometry | Issue |
|---|---|
| Near normal (perpendicular) view | Design transmittance applies best |
| Oblique angle through window | Effective path in optic longer; reflectance up; effective τ down; also target ε may drop |
| Extreme angle | Measurements may be invalid for quantitative work |
Level II practice: center the port, keep the optical axis as square to the window as practical, and note geometry in the report when constraints force compromise.
Windows and the Five Free-Path Parameters
Window τ is additional, not a replacement:
- Still set ε for the target surface (tape on the bus, not “ε of the window”).
- Still set RAT for what the target reflects (the window changes the path but does not eliminate reflection physics at the target).
- Still set T_atm, RH, distance for the air path (follow manufacturer guidance on distance with ports).
- Plus set window transmittance.
Mental model stack
Target emission/reflection → (inside air) → window (τ) → (outside air) → camera optics → detector
Missing any layer misassigns radiance.
Crystal vs Polymer on the Exam
You are not required to memorize every commercial product’s τ. You are required to know:
- Both families need a documented τ for the spectral band
- Impact-resistant polymer ports are common in industry and often have τ significantly below 1
- Crystal ports are chosen for optical performance but still need correct entry and care
- Never use ordinary glass as if it were an IR window in LWIR
Workflow Checklist for Windowed Measurements
- Identify window make/model; retrieve band-specific τ from label, datasheet, or maintenance records.
- Inspect cleanliness and damage; clean if needed and allowed.
- Position for near-normal view; maintain electrical safety distances.
- Enter τ_window on the camera external optics/window setting.
- Enter the five free-path parameters for the actual target.
- Focus on the target plane (not the window dust); confirm spot size on the true target.
- Capture radiometric image; document τ, cleanliness notes, and angle if relevant.
- For trending, use the same port, similar geometry, and same τ practice each visit.
Common Exam Traps
| Trap | Correction |
|---|---|
| Dividing °C by τ | Enter τ; do not scale temperature linearly |
| Setting ε = τ | Different physics |
| Ignoring dirty window | Effective τ drops |
| Measuring through cabinet glass door | Not an IR window |
| Assuming polymer τ = 1 because it looks clear visually | Visible clarity ≠ IR transmittance |
| One τ for all brands without checking band | Use manufacturer band-specific data |
Summary for Recall
IR windows enable safer, repeatable inspection but transmit only a fraction of infrared. Enter that fraction as window transmittance τ in the camera’s radiometric model — for example 0.80 for an 80% window — together with emissivity, RAT, atmosphere, RH, and distance. Never arithmetic-correct displayed temperatures by dividing or adding fixed offsets to “undo” the window. Soiling, damage, condensation, and oblique angles reduce effective transmittance below the clean specification. Level II competence means treating the window as a calibrated optical loss, not as invisible air.
An IR window is specified at 80% transmittance in the camera’s LWIR band. What is the correct quantitative procedure?
Why is dividing a displayed temperature by the window transmittance (T/τ) not a valid correction method?
A previously clean IR window is now coated with dust and fingerprints. What is the most accurate Level II statement?
When measuring a taped high-emissivity spot on a bus through an IR window, which parameter set is appropriate?