The Numbers That Don't Add Up
A luminaire datasheet claims 160 lm/W. The lighting designer trusts this number, calculates that 16 luminaires will deliver 500 lux on the work plane, and signs off the specification. The project is built. The building owner walks in with a lux meter — and measures only 380 lux.
What went wrong? The supplier quoted efficacy from an integrating sphere test. The real luminaire efficacy — measured by a goniophotometer under conditions matching actual installation — is only 128 lm/W. The project is 25% under-lit because the specification was based on inflated data.
This is not a rare edge case. It is one of the most common causes of post-installation disputes in commercial lighting. This article explains what each test actually measures, why suppliers prefer the cheaper method, how the numbers diverge, and how to avoid being misled.
Integrating Sphere: What It Actually Measures
An integrating sphere is a hollow sphere (typically 1–3 metres diameter) coated internally with barium sulphate or PTFE — materials that reflect >95% of incident light diffusely. A luminaire placed inside emits light that bounces repeatedly across the interior until the sphere achieves a spatially uniform radiance field. A spectroradiometer at a detector port then analyses this uniform field.
Primary measurements:
- Correlated Colour Temperature (CCT) — confirming the luminaire is truly 4000K, not drifting to 4500K
- Colour Rendering Index (CRI/Ra) — including R9 for saturated reds
- Chromaticity coordinates and MacAdam ellipse step
- Spectral Power Distribution (SPD)
- Total luminous flux (lumens) — capturing ALL photons emitted in every direction
What it cannot tell you:
- How light is distributed in space (beam angle, intensity at specific angles)
- Whether light actually reaches the target surface
- UGR or glare characteristics
- True luminaire-level system efficacy in real installation
The Integrating Sphere Lumen Trap
Here is the problem many specifiers do not understand: an integrating sphere captures ALL light — including light that would never reach a useful surface in real installation.
What "total flux" really includes:
- Light that hits the back of the luminaire housing and bounces back into the sphere
- Light trapped inside the reflector that never exits the aperture
- Stray light from driver indicator LEDs
- Light emitted at extreme angles (>85°) that in practice hits the ceiling, not the work plane
The orientation problem:
Inside an integrating sphere, a directional luminaire can be oriented to fire directly toward the detector port, maximising the captured flux reading. The same fixture positioned differently inside the same sphere produces a different reading. There is no way to replicate actual installation geometry inside a sphere.
Real example:
A track light with a 24° beam measures 4,100 lumens in an integrating sphere. The same fixture in a goniophotometer — measuring only light that actually exits the optical system in useful directions — produces 3,200 lumens of effective output. That is a 28% inflation. The supplier prints "4,100 lm" on the datasheet and calculates lm/W based on this number.
Why Suppliers Prefer Integrating Sphere Data
The economics are simple — and the incentives are perverse.
| Integrating Sphere | Goniophotometer | |
|---|---|---|
| Equipment cost | €20k–80k | €150k–500k |
| Test time per luminaire | Seconds | 20–60 minutes |
| Lumen result | Higher (inflated) | Lower (real) |
| Produces IES/LDT files | No | Yes |
A factory producing 200 SKUs with multiple beam angle variants faces a choice: invest €300k+ in a goniophotometer and spend weeks testing, or buy a €50k integrating sphere that produces results in seconds — and delivers higher numbers as a bonus.
The temptation is obvious. Test in the cheap, fast sphere. Get higher lm/W numbers. Print those numbers on the datasheet. Higher lm/W on paper wins more tenders against competitors who report honest goniophotometer data.
The customer only discovers the truth after installation, when the lux meter tells a different story. By then, the supplier has been paid, the ceiling is closed, and remediation means adding luminaires at someone else's expense.
Goniophotometer (Dark Room): The Real Performance Test
A goniophotometer measures luminous intensity at thousands of points around a luminaire. Either the fixture rotates on two axes, or a photometer arm orbits around a stationary fixture, in a completely dark room (zero ambient light contamination).
What it measures:
- Luminous intensity at every angle (cd values across a full sphere of measurement points)
- True luminaire luminous flux (only light that actually exits the optic)
- Beam angle (defined as the cone where intensity exceeds 50% of peak)
- Luminaire system efficacy (total useful output ÷ total input power = real lm/W)
- Zonal lumen distribution (how much light goes into each angular zone)
What it produces:
- IES files (.ies) — the standard photometric format in North America and Asia
- LDT/EULUMDAT files (.ldt) — the standard format across Europe
- Polar intensity diagrams and isocandela plots
- UGR tables for glare assessment
Why this is "truth": The goniophotometer measures only photons that leave the luminaire in a specific direction. It cannot be inflated by internal reflections or housing bounce. The measurement represents exactly what a lux meter would detect at a given distance and angle in a real installation.
Why the Gap Causes Real Problems on Site
The common failure path:
- Supplier tests luminaire in integrating sphere → gets 4,100 lm at 38W → claims 108 lm/W
- Lighting designer trusts the datasheet → calculates fixture quantity based on 108 lm/W
- Project is built with the calculated number of luminaires
- Owner measures work plane illuminance → reads 380 lux instead of target 500 lux
- Why? The real luminaire output is 3,200 lm (goniophotometer) → actual efficacy is 84 lm/W
- The installation is 25% under-lit. Adding more luminaires costs money and disrupts the ceiling
Who pays for the gap?
The contractor gets blamed. The designer's reputation suffers. The supplier has already been paid. The owner demands remediation at someone else's cost. Lawyers get involved. Relationships are destroyed — all because of a number that was technically "measured" but practically meaningless.
Key Differences Summary
| Parameter | Integrating Sphere | Goniophotometer (Dark Room) |
|---|---|---|
| Primary purpose | Colour quality verification | Light distribution & real performance |
| CCT accuracy | ✅ Primary measurement | ❌ Not measured |
| CRI / Ra / R9 | ✅ Primary measurement | ❌ Not measured |
| True luminaire lm/W | ❌ Inflated (all photons counted) | ✅ Real system efficacy |
| Beam angle | ❌ Cannot measure | ✅ Primary measurement |
| IES / LDT file output | ❌ Cannot produce | ✅ Primary output |
| UGR data | ❌ Cannot calculate | ✅ Full UGR table |
| DIALux / Relux simulation | ❌ Not usable | ✅ Direct input |
| On-site lux meter match | ❌ Numbers won't match | ✅ Numbers match reality |
| Test cost & speed | Cheap & fast | Expensive & slow |
| Can be "gamed" | ⚠️ Yes (orientation, setup) | ✅ Very difficult to manipulate |
How to Protect Yourself as a Specifier
- Always request lm/W from a goniophotometer report — not a sphere measurement summary
- Cross-check the IES/LDT file: Open the file header — it contains total lumens. Compare this against the datasheet claim. If the datasheet says 4,000 lm but the IES file shows 3,100 lm — the datasheet is using sphere-inflated numbers
- Ask for the test lab name — accredited labs (TÜV, Intertek, SGS, LM-79 certified) follow ISO/CIE standards that minimise gaming
- Request both reports for critical projects — sphere report for colour verification + goniophotometer report for performance verification
- If no IES/LDT file exists — the supplier likely has not done dark room testing. Their efficacy claim is almost certainly sphere-based and inflated
- Demand consistency — the lumen value on the datasheet must match the lumen value in the photometric file. No exceptions.
YNDLUX Dual-Test Protocol
Every YNDLUX luminaire undergoes both tests before release:
- Integrating sphere: CCT tolerance verified within ±100K of nominal (SDCM ≤ 3). CRI Ra ≥ 90 confirmed per production batch. R9 ≥ 50 verified for retail-grade products.
- Goniophotometer: Full IES + LDT files generated for every optical configuration. System lm/W measured at luminaire level (not chip level, not sphere-inflated). Files available for immediate download.
Our rule: The lumen value on the datasheet = the lumen value in the IES/LDT file = what your lux meter reads on site. One number, verified three ways. No sphere inflation. No tricks.
