Wood's Lamp Colour Variation: What Is Acceptable and How to Specify It
Quick answer
Acceptable Wood's lamp colour variation is a written range, not a judgement call: plan on ±5 nm peak wavelength between production lots, ΔE ≤3 on the visible glow against a signed golden sample, and ±15% irradiance at a fixed measuring distance. Specify all three before tooling, because tightening any one of them raises unit cost and MOQ.
- LIGHT BDB's Light & Mood platforms start at 100 pcs standard, 300 pcs white-label and 500–1,000 pcs for ODM custom work.
- Production lead time runs 4–8 weeks from deposit to FOB Shenzhen for most SKUs.
- Sample turnaround is 7–14 days, and the sample fee is refundable against a purchase order.
- Every unit ships under a 500-day limited warranty, one of the longer terms in this category.
- Ask for a golden sample signed by both parties; without one, any ΔE tolerance you write is unenforceable.
- Most LED suppliers bin 365 nm UV-A parts in 5 nm bins, so a ±5 nm lot spread is industry-typical.
- LIGHT BDB holds ISO 9001, ISO 14001, BSCI (amfori), UL registered factory and Sony Green Partner status.
- T/T 30/70 and L/C at sight are both accepted, FOB Shenzhen, with your nominated forwarder.
What colour variation is actually acceptable in a Wood's lamp order?
Acceptable variation is a three-part written range: ±5 nm peak wavelength between lots, ΔE ≤3 on the visible glow, and ±15% irradiance at a fixed distance.
Everything outside those three numbers is a negotiation, not a specification. The reason buyers get burned in this category is that they judge a Wood's lamp by what it looks like on a desk in a bright room, then discover six months later that half the shipment glows a slightly different violet. UV-A output is largely invisible; the violet you see is the small visible tail of the emission plus whatever leaks through the filter. Your eye is a poor instrument for measuring it and a worse instrument for comparing two units side by side.
A note on terminology: Wood's lamp versus wooden lamp
This guide covers the Wood's lamp — the UV-A fluorescence inspection light used in beauty, material and authenticity inspection work. If you are sourcing a wooden decorative lamp instead, the discipline is the same but the measurement target changes: you are matching stain and grain on the housing rather than spectral output, and the practical tolerance is ΔE ≤2 against a master plaque under D65 lighting.
The three layers buyers actually argue about
- Spectral layer. Peak wavelength, half-bandwidth and filter cut-off. This is what the lamp does. It is invisible to the buyer and measurable only with a spectroradiometer.
- Visible appearance layer. The colour of the glow on the shelf, in photos and in a salon. This is what gets complained about, and it is driven mostly by visible leakage through the filter.
- Cosmetic layer. Housing, trim, button and diffuser colour. This is ordinary plastics colour control: ΔE against a master, gloss units, texture.
Write a separate tolerance for each layer. Mixing them into one "colour must match the sample" sentence is the single most common cause of failed first articles in this product family.
What actually drives Wood's lamp colour variation between units?
Five things drive it: LED die binning, filter-glass batch, driver current, thermal design and mechanical assembly geometry.
LED binning is the biggest lever
UV-A LEDs are sold in wavelength bins, and most suppliers in this category bin 365 nm parts in 5 nm steps. Two reels from the same purchase order can sit at opposite ends of a bin. If your specification does not name a bin, the factory will buy whatever is available and cheapest that week, and your lot-to-lot shift is already baked in before assembly starts. Buyers who want tighter consistency typically pay a premium for narrow-bin or single-bin supply, and the MOQ the LED distributor imposes then flows down to your order quantity.
The Wood's glass filter is the second
The filter is what separates a deep, clean violet glow from a washed-out bluish-white one. Filter glass and coated alternatives vary batch to batch in transmission above 420 nm. A good filter blocks the vast majority of visible light and lets the UV-A through; a cheap one leaks a visible blue-white component that makes the lamp look brighter and less "Wood's lamp" in photographs. Two lamps with identical LEDs and different filter batches will look different to a customer, which is why the filter belongs in the specification, not just the bill of materials.
Driver, thermals and assembly geometry
- Drive current tolerance. A driver running ±5% current moves both irradiance and, slightly, the emission peak.
- Thermal design. UV-A LEDs shift wavelength upward as junction temperature rises. A housing that runs 15 °C hotter at steady state will read differently after ten minutes of warm-up.
- LED seating depth and reflector geometry. A fraction of a millimetre in seating depth changes beam uniformity, which changes how the glow reads to the eye even when the spectrum is identical.
- Lens and diffuser material. Some optical plastics yellow measurably within a year under UV exposure, shifting the visible tint long after the goods have shipped.
That last point matters more than most buyers expect. Ask what the diffuser is made of, and ask for a UV ageing data point rather than an assurance.
How do I write a colour spec the factory can actually measure?
Write it as a parameter, a method, a geometry, a sample size and a pass/fail rule — never as "same as last time".
- Name the parameter. Peak wavelength, FWHM, filter transmission above 420 nm, chromaticity coordinates, irradiance at a stated distance, housing ΔE.
- Name the instrument. Spectroradiometer for spectral and chromaticity work, UV radiometer for irradiance, spectrophotometer for plastics. A phone camera is not an instrument.
- Fix the geometry. Distance from lamp to sensor, angle, ambient temperature, dark-room conditions. A measurement at 10 cm and one at 38 cm are not comparable and never will be.
- Fix the warm-up. Ten minutes at 25 °C ambient is a workable default; UV-A output drifts in the first few minutes and settles after that.
- Fix the sample size and the rule. How many units per lot, and what happens if one fails. Reference an AQL plan rather than "we will check a few".
- Attach the golden sample. A physical unit, signed and dated by both parties, stored dark and cool, with a spectral scan on file to back it up.
The table below shows the parameters worth putting in a purchase order and the tolerance ranges that are industry-typical for LED and luminaire factories in this category. They are starting points for negotiation, not fixed rules, and every one of them can be tightened at a cost.
| Parameter | Test method | Industry-typical acceptance | What it controls |
|---|---|---|---|
| Peak wavelength (λp) | Spectroradiometer, 25 °C, 10-minute warm-up | 365 ±5 nm; ≤5 nm spread within one lot | Whether units are genuinely UV-A and consistent in use |
| Spectral half-bandwidth (FWHM) | Same scan | ≤15–20 nm | Purity of the UV-A peak versus broad spectral spill |
| Visible leakage above 420 nm | Filter transmission scan, 380–450 nm | ≥90% blocking for good Wood's glass | Deep-violet glow versus washed-out bluish-white |
| Visible glow chromaticity | Calibrated spectroradiometer, CIE 1931 x/y | ΔE ≤3 (CIE2000) or inside a 3-step MacAdam ellipse versus golden sample | Shelf and photo consistency across lots |
| Irradiance at reference distance | UV radiometer at the agreed distance and angle | ±15% unit-to-unit | Perceived brightness and beam strength |
| Housing and trim colour | Spectrophotometer, 60° geometry, D65 | ΔE ≤2 versus master plaque; gloss ±5 GU | Cosmetic match to your brand system |
| Ageing drift | 1,000-hour burn-in on a sample of five units | ≤5% irradiance drop; ≤2 nm peak shift | What the end customer sees in month six |
| Lot documentation | Spectral scan and bin-lot record per batch | Report attached to each shipment | Dispute resolution and reorder continuity |
One practical note on sampling: most factories in this category inspect visual and cosmetic attributes to AQL 2.5 for major defects and 4.0 for minor, and treat spectral parameters as a lot-level test on a handful of units rather than a 100% screen. If your product is premium-priced, buy a 100% irradiance check — it is fast, it is cheap, and it catches the driver outliers that a five-piece sample misses.
Which standards and safety points belong in a Wood's lamp spec?
Reference IEC 62471 photobiological safety, RoHS and REACH, plus your destination market's electrical approvals — UV-A lamps typically sit in Risk Group 2 or 3.
Risk group classification drives your labelling, your warning text, your packaging and, in some markets, your listing documentation. Build it into the specification early, because warning artwork changes are slow and expensive once tooling is cut. Practical items worth writing down:
- Eye-hazard wording on the product and the insert, plus supplier-goggles advice where the risk group calls for it.
- Exposure time guidance and an auto-off timer if the lamp is handheld or consumer-facing.
- UV-blocking diffuser options for markets that restrict consumer UV exposure.
- Documented risk-group test report from a recognised lab, referenced by report number in the PO.
On the factory side, LIGHT BDB holds ISO 9001, ISO 14001, BSCI (amfori), UL registered factory and Sony Green Partner status, and we run a 10,000+ sqm component plant in Panyu with our exclusive strategic manufacturing partner Fuli Electronics, a 3,000 sqm finished-product assembly base in Shenzhen and a 5,000+ pcs/day overflow line in Yangon. A UL-registered factory shortens the paperwork path for UL or ETL submissions, though the listing itself applies to the finished product under your brand, not to us.
Be clear internally about what you are selling. LIGHT BDB is a wellness-device ODM: the lamps we build are consumer light products, and nothing in the marketing should carry diagnostic or disease claims in any market. If your business plan depends on regulated healthcare positioning, this category and this supply chain are the wrong fit, and it is better to know that before the tooling invoice.
What MOQ, lead time and warranty apply to a Wood's lamp build?
On our Light & Mood platforms, MOQ is 100 pcs standard, 300 pcs white-label and 500–1,000 pcs for ODM custom, with 4–8 week lead times.
- 100 pcs — standard platform, your logo on a catalogue build.
- 300 pcs — white-label with your packaging, artwork and accessories.
- 500–1,000 pcs — ODM custom: new housing, custom optics, your chosen bin and filter.
- Lead time — 4–8 weeks from deposit to FOB Shenzhen for most SKUs.
- Samples — 7–14 days, fee refundable against the purchase order.
- Warranty — 500-day limited warranty, one of the longest in the category.
- Terms — T/T 30/70 or L/C at sight, FOB Shenzhen, your nominated forwarder.
Expect the tighter colour windows to pull against these numbers. Narrow-bin LED supply usually comes with a distributor minimum that can exceed your production quantity, and a custom filter run adds weeks. If your launch quantity is 100 pcs, choose a standard platform and accept a wider but documented tolerance; if colour consistency is the core of your brand promise, plan for the ODM tier from the start.
How do I keep colour consistent across reorders?
Retain a signed golden sample, a spectral scan per lot and a bin-lot record, then repeat the identical measurement on every reorder.
The failure mode is predictable: the second order arrives eight months later, the LED bin has changed, the filter supplier has changed, and nobody can prove what "matching" meant. Three habits prevent it.
- Keep three golden samples. One at the factory, one at your office, one in a cool dark cabinet as the tie-breaker. Signed and dated.
- Store the digital record. The spectral scan, the chromaticity coordinates, the irradiance figure and the bin-lot number for every shipment, kept for at least three years.
- Write a change-notification clause. Any change to LED bin, filter supplier, driver or diffuser material requires your written approval before production, with a fresh sample.
Also decide in advance what happens when a reorder is "close but not identical". A workable rule: within ΔE 3 is acceptable and ships; between ΔE 3 and ΔE 5 is a commercial decision you make jointly; above ΔE 5 is a reject. Writing that ladder into the agreement now saves a very awkward conversation later.
What are the honest trade-offs in specifying Wood's lamp colour?
Tighter colour windows cost more, add lead time and raise MOQ, and no factory can promise zero variation.
Three trade-offs deserve to be named plainly. First, measurement costs money: a calibrated spectroradiometer report per lot is not free, and buyers who refuse to pay for verification end up arguing over photographs instead. Second, purity versus perceived brightness — a well-filtered 365 nm lamp looks dimmer to the eye than a leaky 395 nm lamp, and some buyers reject the technically better unit because it "looks less bright" on a video call. Third, cosmetics versus volume — a ΔE ≤1.5 housing tolerance on a 100-piece run is not realistic from any factory, whatever the quotation says.
There is also a genuine limit on the supply side. LED output shifts with age, filters vary by melt batch, and the measurement itself has uncertainty. A specification that assumes perfect repeatability will fail on the third order no matter who builds it. The realistic goal is a documented, verifiable, consistently applied tolerance that your customer cannot see across units in normal use. That is achievable. Perfect uniformity is not, and any partner who tells you otherwise is selling you a problem.