WELL v2 Lighting Requirements: A Supplier Checklist

WELL v2's Light concept contains nine features: two preconditions and seven optimizations worth eighteen points. Only two of those features are closed out by a measurement taken in the finished space — L02 Visual Lighting Design and L03 Circadian Lighting Design — plus one optional route inside a third, L01's circadian lighting route, which borrows the L03 threshold. Everything else is settled by documents: test reports, calculation sheets, control certificates, operation schedules. That single fact decides whether a lighting package survives a WELL project, and it is the reason a supplier's first deliverable is not a luminaire but a document set that has to exist before an assessor visits. This article lists the nine features with their current thresholds, identifies the document each one is verified by, and turns the result into a supplier checklist you can attach to the purchase order. Every threshold below was read from the standard's own feature text on 2026-09-30, and every internal figure is a figure SUMLINK already publishes.

Key Facts

1. WELL v2's Light concept has nine features — two preconditions (L01 Light Exposure, L02 Visual Lighting Design) and seven optimizations worth 18 points in total (L03: 3, L04: 2, L05: 4, L06: 2, L07: 1, L08: 3, L09: 3).
2. L03 Circadian Lighting Design awards 1 point at ≥150 EML [136 melanopic EDI] and 3 points at ≥275 EML [250 melanopic EDI], maintained for at least four hours beginning no later than noon, measured on the vertical plane at eye level at 18 in (45 cm) above the work-plane. Projects claiming enhanced daylight (L05 Part 1 or L06 Part 1) may use the lower pair — 120 and 180 EML.
3. L08 Part 2 (2 points) accepts any one of three flicker routes: California Title 24 "reduced flicker operation" tested per Joint Appendix JA-10; IEEE 1789-2015 recommended practice 1, 2 or 3; or PstLM ≤ 1.0 with SVM ≤ 0.6 — applied to the luminaires and their controls.
4. Where tunable white is used, L08 Part 1 requires the colour-rendering route to hold at 1000 K intervals from the lower end (minimum 2700 K) to the higher end (maximum 5000 K) — not only at the two endpoints.
5. L04 Electric Light Glare Control accepts UGR ≤ 19 (on either the luminaire route or the space route), or luminance ≤ 6,000 cd/m² at any angle between 45° and 90° from nadir on the luminaire route; the luminaire route is assessed at "light output representative of regular use conditions" — the operating level, not a full-output datasheet.

Terms and units used below. EML = equivalent melanopic lux; melanopic EDI (m-EDI) = melanopic equivalent daylight illuminance — the two are ways of writing the same melanopic quantity, which is why the standard states every threshold twice. SVM = stroboscopic visibility measure and PstLM = short-term flicker severity, the two flicker metrics the standard names. UGR = unified glare rating; sDA = spatial daylight autonomy; VLT = visible light transmittance of the glazing; SDCM = standard deviation of colour matching (MacAdam step). fc = foot-candle (1 fc ≈ 10.76 lx); ft and ft² are feet and square feet; cd/m² is luminance; Ta is ambient temperature. The expansions are ours; the symbols are as the standard and the cited standards write them.

The nine features, and the column that is actually a work order

The standard states, for every feature, how the achievement is verified. That column is the supplier's work order.

Two observations follow from that table, and both change how a lighting package is bought.

First, the assessment is mostly a paperwork exercise. Two features are verified by a test in the finished space — L02 (performance test plus an individual technical document) and L03 (performance test alone) — and L01 adds a third, because its fourth route, circadian lighting design, is verified by a performance test: it borrows L03's Tier 1 threshold. Everything else in the Light concept is an audited document, a narrative or a schedule.

Second, "Technical Document (Audited)" is not a promise about the product. It is a description of the paperwork the project must hold. If the luminaire's LM-79 report, its UGR table, its spectral data or its driver certificate cannot be produced as a signed document at the moment the assessment is scheduled, the corresponding points are not at risk — they are unavailable, regardless of how the installation measures.

The two preconditions: the gate, and the one route a lighting supplier can carry alone

L01 Light Exposure requires indoor light exposure through daylight and electric light strategies. The feature offers four alternative routes.

Options 1 to 3 are architecture. Option 4 is electric light, and it is the only one a lighting supplier can carry without waiting on a facade drawing: it converts the precondition into a measured melanopic target in the space. It is also the strictest of the four in effort, because it inherits L03's measurement protocol.

L02 Visual Lighting Design requires appropriate illuminance on the work planes for occupants of all age groups, as required by the tasks performed.

In SI units the standard states the same levels as 320 lux at the task surface and 110 lux at floor level and at task surface — the feature's unit toggle reads "30 fc | 320 lux" and "10 fc | 110 lux". A strict conversion gives 323 lx for 30 fc and 108 lx for 10 fc, so project documents carry both numbers; the standard's own figure is the one to quote.

The supplier's part in L02 is narrower than it looks, and it is the part that is most often missing: an illuminance calculation per space, the named reference guideline, the age assumption used, and the light-loss maintenance factor. The performance test then measures the room, not the datasheet.

L03 Circadian Lighting Design: a threshold, a window and a plane

L03 is where the Light concept stops being about lux and starts being about time. Three details carry the feature.

Three conditions sit on top of the thresholds, and each one is a supplier question:

• Duration and timing. The levels must be achieved for at least four hours, beginning by noon at the latest. A fixture that can reach 275 EML is not compliant; a schedule that holds it across the morning window is.
• Height. The target is measured at 18 in (45 cm) above the work-plane — the standard's own metric equivalent — not on the desk surface.
• Plane. The light levels are achieved on the vertical plane at eye level, to simulate the light entering the eye.

The verification is a performance test, which means the project has to be able to state, in advance and in writing: the spectral data behind each fixture's melanopic output, the CCT and dim level at each schedule node, the illuminance the node is expected to produce, and the m-EDI that follows from it. A supplier that can only quote one melanopic number for a product — a single CCT, usually at full output — cannot support this feature, because compliance is a time series rather than a point value.

What that looks like in practice is published on the BY001 series page: the BY001 series carries a per-CCT melanopic table across 2700 K to 6500 K, with an M/P ratio range of 0.45 to 1.06 and a melanopic EDI column from 225 at 2700 K to 530 at 6500 K at 500 lx. Every row of that table is labelled a design target — an engineering design value from the LED and driver platform, not an independent laboratory result. That labelling is the standard a WELL project should demand from every supplier, because the difference between a design target and a measured value is exactly the difference between a document the assessor can audit and a document they cannot.

L04 Electric Light Glare Control: measured at the output you actually run

L04 is worth 2 points and is verified by one audited technical document. It offers two routes.

The qualifier that decides most disputes sits on Option 1, the luminaire route: the assessment is made when measured at light output representative of regular use conditions. Option 2, the space route, requires a UGR classification of 19 or lower for every regularly occupied space and names no measurement condition; the luminance sub-route sits inside Option 1 and inherits the qualifier with it. A UGR figure taken at full output tells an assessor what the room does at full output. On a tunable-white installation that dims to 10% in the evening, the glare assessment has to follow the operating points the program actually uses, and the exclusion list is explicit, so a wall-washer cannot be counted as evidence for the rest of the ceiling.

Suppliers can meet this with either artefact. A UGR table for the luminaire in the specified ceiling layout (4H8H transverse viewing is the conventional basis) is the usual one. Luminance-versus-angle data is the alternative, and it is the more robust of the two on a dimmed system, because it lets the designer re-evaluate the installation at any output. SUMLINK publishes UGR classifications for two lines in its catalogue: `< 19` on the BY001 series, where the value carries an explicit "Design target" label, and `< 13` across all CCT states on the Linan 2.0 recessed spotlight, where the catalogue index lists the figure with no label at all. An unlabelled catalogue figure is not assessment evidence, so treat both as unconfirmed until a signed report is attached.

L05 and L06: daylight, and the boundary of the lighting supplier's scope

L05 Daylight Design Strategies (4 points) and L06 Daylight Simulation (2 points) are the two largest blocks in the Light concept that a lighting supplier cannot deliver. They are verified by documents rather than by luminaire test data — L05 Part 1 and L06 Part 1 as audited technical documents, L05 Part 2 by on-site photographs plus a policy or operations schedule — and every one of those documents belongs to the architecture. They belong in this checklist for one reason: they change the melanopic target the lighting package has to hit.

The link back to the lighting package is the "enhanced daylight" column in L03. A project that meets L05 Part 1 or L06 Part 1 can claim 1 point from 120 EML instead of 150, and 3 points from 180 EML instead of 275. That is not a small difference: at the top tier it removes roughly a third of the required melanopic output, and it is the difference between a package that can reach the target with a tuned white ambient system and one that needs dedicated high-melanopic fixtures.

Which means the first question a lighting supplier should ask on a WELL project is not about the products. It is: is the project claiming the enhanced-daylight pathway, and which of L05 Part 1 or L06 Part 1 is it claiming it with? If the answer changes later, the lighting schedule changes with it. The supplier's own deliverables in this pair are narrow and specific — IES photometric files for the simulation, dimming curves and driver linearity for daylight-linked control, and confirmation of the CCT and dim nodes the control system will hold when daylight is contributing.

L07 Visual Balance: one point, four parameters, and a fade time

L07 is worth 1 point and is verified by a professional narrative. Part 1 offers two routes.

Two of the four parameters are product properties rather than design choices, and they are the two that get missed.

The ±200 K consistency requirement is a binning and driver-accuracy statement. Two luminaires of the same model from different production bins, or on two drivers with different tuning curves, can sit more than 200 K apart at the same nominal setting — and the feature is assessed per room, per point in time. A supplier that publishes SDCM or a colour-shift figure but no CCT tolerance should be asked for the tolerance directly.

The 10-minute change requirement quietly bans the dramatic transition. A program that sweeps from 5000 K to 2700 K across two minutes photographs well and is a professional-narrative problem; the transition time is set in the gateway or driver, so it is a commissioning setting rather than a purchase. Write it into the commissioning checklist instead of the specification table.

L08 Electric Light Quality: three points, and two clauses that catch suppliers out

L08 is worth 3 points across two parts, both verified by an audited technical document. It is the feature most directly aimed at the luminaire.

The tuning-range clause is the first trap. A product that publishes CRI 90 at 4000 K and nothing else has characterised a single point on a range the feature makes the project read in 1000 K steps, from a lower bound of 2700 K to an upper bound of 5000 K. If the catalogue states CRI 90 as the platform standard across the range, that is a claim about the range, and an assessor will read it as one. A supplier should be able to hand over colour data at the intervals the standard names, or state plainly that the range is not characterised. (The standard fixes the bounds and the step size but publishes no list of assessment nodes, so the node list is ours to propose — and a supplier should be able to fill it in.)

The controls clause is the second trap. Part 2 applies to the luminaire and its controls. Flicker behaviour is a property of the driver and dimmer, not the LED board, and it changes with dim level: a driver that is clean at full output can produce visible modulation in the last 10% of its range. A flicker statement that covers only full load therefore documents a condition the project may never operate in.

That distinction is visible in published numbers if you look for it. SUMLINK's BY001 page quotes the EU Ecodesign Regulation 2019/2020 limit as SVM ≤ 0.4 at full load, and publishes SVM < 0.15 with percent flicker < 1% across the dimming range as its own design targets — two different scopes, one regulatory and one operating. The WELL threshold in Part 2 (PstLM ≤ 1.0 and SVM ≤ 0.6) is a third scope again. The three are not interchangeable, and the report a WELL project needs is the one that covers the dim levels the program uses.

In practice, the deliverable for Part 2 is a flicker test report that names the driver and the control device tested, states the dim levels, and gives PstLM and SVM per level — or the JA-10 classification, or the IEEE 1789 practice number, with the measurement basis identified. Format matters less than traceability: the report has to describe a fixture-plus-driver-plus-dimmer assembly that matches what the project buys.

L09 Occupant Lighting Control: the points that leave the room

L09 is worth 3 points, split between controllability and a supply commitment that most lighting quotes omit entirely.

Part 1 is a zone-count exercise, and it is decided by the layout drawing, not by the luminaire. Two numbers dominate: 320 ft² per zone for the second point, and three levels or scenes per zone. Both are cheap to design in at tender stage and expensive to retrofit, because adding zones to a completed DALI or wireless installation means re-addressing devices and re-commissioning.

Part 2 is unusual in the Light concept: it is a purchasing policy with an eight-week service standard, verified by policy or operations schedule, and it is worth a full point. A supplier that ships tunable-white ambient lighting and no task-light option has already forfeited that point — and the requirement that the light-emitting element is not visible under intended use rules out most of the office task lights bought on price.

The verifications split accordingly: the zone table is an audited technical document; the control-system behaviour is a professional narrative; the supplemental-lighting availability is a policy or operations schedule; and the characteristics of the supplemental fixtures themselves — independent control, adjustable location, non-visible emitter — are an audited technical document. Four verification statements, four different owners inside the project, which is why L09 is the feature most often only partly achieved.

The checklist

This is the article's deliverable. Each row is a document; the third column is the feature that will not be awarded without it. Issue it with the purchase order, and ask for dated files rather than availability.

Rows 3, 4 and 7 are the ones most often missing altogether, and they are the three that decide the electrical side of the Light concept. Row 13 is not a luminaire document at all, and it is the row most likely to change what the supplier has to deliver: if the project cannot claim L05 Part 1 or L06 Part 1, the L03 target moves from 180 EML to 275 EML for the third point.

What SUMLINK publishes today, and what has to be requested

A checklist is only useful if it says where the project stands against it. SUMLINK is a lighting strategy supplier and does not manufacture products, so its published position is about how figures are labelled and how manufacturer documents are handed over.

What is published without a request. The catalogue index leads with the parameters the about page calls "the parameters that actually predict HCL outcomes" — M/P ratio, SVM, Δu′v′, CCT tuning range — alongside the standard photometric and electrical fields, and the about page states that full IES files download without registration. The BY001 series page publishes its M/P range (0.45–1.06) and a melanopic EDI column from 225 at 2700 K to 530 at 6500 K at 500 lx. It also publishes percent flicker below 1% across the dimming range, SVM below 0.15, Δu′v′ ≤ 0.004 at 6,000 h and Ta 25 °C, SDCM ≤ 3-step, IES TM-30-20 Rf 88 / Rg 99, and UGR below 19.

What the labels say. Every one of those figures is marked Design target on the page, and the page states the policy plainly: values marked Design target are engineering design values from the driver and LED platform specification, not yet confirmed by independent laboratory report; verified values are backed by a test report available on request. The catalogue standard, stated on SUMLINK's about page, is tunable white at CRI 90 with R9 ≥ 50.

What has to be requested. The Download rows for datasheets, IES/LDT files, installation and DALI wiring diagrams, BIM/Revit families, test reports for the verified metrics, and warranty and compliance documents are all marked "available on request". Read against the checklist above, that puts the test-level rows — LM-79 reports (row 1), UGR or luminance data (row 5), colour data across the tuning range (row 6) and flicker at operating levels (row 7) — in the request-issued category. Request-issued is a different problem from missing: rows 3, 4 and 7 are the ones that tend not to be in a supplier's file at all.

The honest gap. No published catalogue row claims an independent laboratory verification of its photometric or human-factor figures. What the BY001 series page claims is narrower, and worth reading in its own words: values marked "Design target" are "engineering design values from the driver and LED platform specification, not yet confirmed by independent laboratory report", while "verified values are backed by a test report available on request". That label lives on the series pages; the catalogue index lists figures such as `< 13` UGR or CRI 97+ with no label at all. A WELL project's technical documents therefore still have to come from the manufacturer, signed and dated, and the supplier's job is to obtain them, check that the model, the driver and the controls in the report match what is being sold, and hand them over with the HCL specification library reference.

Six clauses to write into the specification

Copy these into the lighting specification, or into the purchase order as compliance conditions. Each one closes a gap that the standard's verification column opens.

1. Melanopic nodes. "For each CCT and dim node in the approved schedule, the supplier shall provide the spectral power distribution, the melanopic daylight efficacy ratio, and the resulting melanopic EDI at the specified maintained illuminance, calculated in accordance with CIE S 026:2018 (SUMLINK's specified calculation basis; the WELL feature text's own melanopic reference is PD CEN/TR 16791:2017). Values shall be stated per node, not as a single product figure."
2. Full-range colour data. "Where tunable white luminaires are supplied, colour rendering data shall be provided at 1000 K intervals from the lower end of the tuning range (minimum 2700 K) to the higher end (maximum 5000 K), stating CRI (Ra) and R9, or TM-30 Rf, Rg and Rcs,h1."
3. Flicker at operating levels. "Flicker performance shall be reported for the luminaire together with its specified driver and control device, at each dim level used by the approved schedule, stating PstLM and SVM, or the Joint Appendix JA-10 reduced-flicker classification, or the applicable IEEE 1789-2015 practice."
4. Glare at regular-use output. "UGR classification and luminance-versus-angle data shall be provided for the luminaires in their specified ceiling layout and at light output representative of regular use conditions."
5. Control completeness. "The control system shall provide a minimum of three lighting levels or scenes per zone, occupant control of immediate lighting within the zone, CCT or colour or distribution change by group or preset scene, and separate control of lighting serving presentation or projection walls. Zone count shall be stated as floor area or occupant count per zone."
6. Document delivery dates. "All technical documents required for verification of the project's WELL v2 Light features shall be delivered no later than four weeks before the scheduled assessment date, and shall identify the model, driver and control device supplied."

The sixth clause is the one that changes outcomes. Every document above can exist and still arrive after the assessment window has closed, at which point availability is worth nothing.

The point math, and what to check against your own scorecard

Adding the feature values gives 18 optimization points in the Light concept: L03 3, L04 2, L05 4, L06 2, L07 1, L08 3, L09 3. The two preconditions carry no points and are conditions of entry.

Sorted by who has to produce the evidence, the 18 points fall into three groups:

• 9 points sit on luminaire and control documentation — L03 (3), L08 (3) and L09 (3). These are the points a lighting supplier materially controls, and all three are verified by supplier documents: L03 by a performance test, L08 by an audited technical document, and L09 by an audited technical document, a professional narrative and a policy or operations schedule.
• 3 points depend on electrical design documents — L04 (2) and L07 (1), where the artefacts are a UGR or luminance assessment and a professional narrative.
• 6 points are daylight and architecture-led — L05 (4) and L06 (2), plus three of the four routes into the L01 precondition.

On a WELL v2 office or hotel project, half of the Light concept's optimization points are decided by paperwork that the lighting package is expected to produce, and the biggest single lever on the melanopic target — the enhanced-daylight pathway — is decided by a facade drawing.

If your supplier cannot date-stamp a document against each row of the checklist in §D.9, the project is buying luminaires, not a WELL v2 lighting strategy.

Frequently Asked Questions

What documents does a WELL v2 lighting project need from the luminaire supplier?

Following the standard's verification column: an LM-79 test report and IES or LDT photometric files for L02 and L04; spectral power distribution with melanopic DER per CCT node for L03; a UGR table or luminance-versus-angle data at operating output for L04; colour data at 1000 K intervals from 2700 K to 5000 K for L08 Part 1; a flicker report naming the driver and the control device with PstLM and SVM per dim level for L08 Part 2; and control documentation — DALI-2, Matter or Casambi certificates, the scene or level list, and the personal-control interface — for L09.

What melanopic level does WELL v2 L03 require?

L03 Circadian Lighting Design awards 1 point at ≥ 150 EML [136 melanopic EDI] and 3 points at ≥ 275 EML [250 melanopic EDI]. Projects that meet enhanced daylight through L05 Part 1 or L06 Part 1 may use the lower pair: ≥ 120 EML [109 melanopic EDI] for 1 point and ≥ 180 EML [163 melanopic EDI] for 3 points. The levels must be maintained for at least four hours beginning no later than noon, measured on the vertical plane at eye level at 18 in (45 cm) above the work-plane.

What flicker limits does WELL v2 accept?

L08 Part 2, worth 2 points, accepts any one of three routes: classification as "reduced flicker operation" under California Title 24 when tested per Joint Appendix JA-10; IEEE 1789-2015 recommended practice 1, 2 or 3; or PstLM ≤ 1.0 together with SVM ≤ 0.6. The requirement applies to the luminaires and their controls, so the evidence has to cover the dim levels the project operates at rather than full output alone.

Does WELL v2 require colour rendering at every CCT on a tunable-white installation?

Yes. L08 Part 1 requires the chosen colour route — CRI (Ra) ≥ 90, or CRI (Ra) ≥ 80 with R9 ≥ 50, or IES TM-30 P1 with Rf ≥ 78, Rg ≥ 95 and Rcs,h1 between −1% and 15% — to be met at 1000 K intervals from the lower end of the tuning range (minimum 2700 K) to the higher end (maximum 5000 K).

Are the two WELL v2 Light preconditions harder than the 18 optimization points?

They are different in kind. L01 Light Exposure and L02 Visual Lighting Design must be satisfied regardless of points, so they act as a gate; only the L01 circadian-lighting route and the L02 performance test involve a lighting measurement, and the circadian route simply borrows L03 Tier 1. The 18 optimization points are where the electrical documentation concentrates: L03, L08 and L09 together carry 9 of the 18 points and are verified by supplier documents.

Request BOM & Spec Sheets

Tell us which WELL v2 Light features your project is claiming and we will issue the document set against the checklist. Request BOM & Spec Sheets →