Breadcrumb: Home > Applications > Semiconductor Cleanroom
Updated:2026-11-17
Views: The difference between a surface inspection lamp and ordinary lighting is not brightness but how the light strikes the workpiece. White, yellow-green and yellow light each have a role, and with the wrong spectrum or the wrong angle a micron-scale scratch disappears into the reflection. Illuminance alone does not decide whether a defect is visible; the angle and the uniformity of the spot do. This article answers 18 common questions from semiconductor cover glass, glass, polished metal, plastic paintwork and cleanroom visual inspection in five groups - principle, interpretation, workstation, power and maintenance, and acceptance - and compares three models against their published parameters.
The difference between a surface inspection lamp and ordinary lighting is not brightness but how the light strikes the workpiece. White, yellow-green and yellow light each have a role, and with the wrong spectrum or the wrong illumination angle a micron-scale scratch disappears into the reflection. The 18 questions most often asked about semiconductor cover glass, glass, polished metal, plastic paintwork and cleanroom visual inspection are grouped below into five blocks - principle, interpretation, workstation, power and maintenance, and acceptance - with selection advice based on the actual parameters of the LUYOR-3360G yellow-green inspection torch, the LUYOR-3325G portable yellow-green surface inspection lamp and the LUYOR-3318 series bench surface inspection lamp.
Ordinary lighting solves "illuminating" and nothing more: even, bright enough, not glaring. A surface inspection lamp solves "making the defect visible": direction and contrast. It uses directional condenser optics to gather the light into a beam of consistent direction and throws it obliquely onto the workpiece, so micron-scale scratches and particles scatter light into the eye while the background stays dark, and the contrast is amplified.
Three differences matter in real work. Beam direction: a near-collimated beam at a low angle picks out particles and shallow scratches more readily than diffuse light. Glare control: a mirror-finish workpiece under an ordinary luminaire goes white across the whole surface and the defect drowns in the reflection. Spot uniformity and freedom from flicker: the former decides whether the criterion is the same at every position on the part, the latter whether long visual inspection misses defects that appear and disappear. Together these are the real gap between a dedicated inspection lamp and an ordinary light.
The role of each spectrum follows from what creates the contrast for the defect you want to see; it is not a difference in brightness.
| Spectrum | Band | Strength | Typical workpieces |
|---|---|---|---|
| Yellow-green | 500-600nm | In the high-sensitivity region of the eye's visual response, so it looks brighter at the same power; the narrow band suppresses colour interference, giving clean contrast on particles, micro-scratches, dents and ripple | Wafers, glass cover plates, lenses, polished metal |
| White | 400-760nm | True colour rendering, so a defect and a colour difference can be judged together | Colour deviation, contamination, uneven coating, full visual inspection of machined parts |
| Yellow | 530-670nm | Contains no blue-violet or ultraviolet component, so it does not expose photoresist | Photoresist yellow rooms, photosensitive material areas |
In the 3360 series, for example, the three versions are identical in body and battery and differ in output spectrum alone: 3360G is 500-600nm yellow-green, about 280,000 lux in a 5 cm spot at 40 cm; 3360W is 400-760nm white at about 130,000 lux; 3360Y is 530-670nm yellow at about 70,000 lux. The narrower the band, the more is filtered out, and the higher the visible intensity per unit area at the same setting - which is why the G version shows particles more readily.
Where a large area has to be swept rather than held in the hand, look at the 3325G: 36W, 570-580nm, 900g, 180×200mm, carried along the machine on foot or set on a bench bracket for a fixed station. For long full inspection at a bench, use the 3318: 24W with a diffuser plate for soft area light, cold white 6000K at about 59000 lux and warm white 3000K at about 49000 lux (at 30cm), with an optional 500-600nm green filter that shows scratches and particles on glass and coated surfaces more cleanly.
No. The criteria are entirely different: a visible-light inspection lamp looks for geometric defects - scratches, particles, pits, press marks - and relies on the light-dark contrast produced when light is scattered; a UV blacklight looks for materials that fluoresce - oil, cleaning agent residue, fluorescent penetrant - and relies on the material emitting light of its own when excited at 365nm.
Yellow-green is visible light and cannot excite fluorescence; an oil film under it merely looks slightly darker. Conversely, non-luminous particles such as metal swarf, dust and glass fragments are black under a UV lamp, and visible parallel light is what picks them out. The common practice in a cleanroom is to have both: a UV lamp to confirm organic residue, then yellow-green light to sweep for particles and scratches. For the UV side see the 365nm fluorescent inspection blacklight section; for clean-area rounds compare the MNT-365A cleanroom particle inspection torch.
Flicker has two consequences, and both turn directly into missed and misjudged defects. Persistence of vision: under a rippling brightness the eye sees a fine scratch or particle appear and disappear, and a defect that happens to fall in a dark cycle as you sweep past is missed. Fatigue: working for hours under fluctuating light increases the load on the eye's accommodation, and the criterion loosens as the shift goes on.
There is also a scenario that is easily overlooked - photographic evidence. A flickering source photographs as bands of light and dark, and the defect is then read as a "banding defect", so the record cannot serve as evidence for re-inspection. The 3325G uses a constant-output, flicker-free design, which is exactly what long continuous visual inspection and photographic records need; the 3360 series is likewise constant-output with no warm-up, so a press of the button starts the inspection.
They can, and this is where yellow-green light is at its most useful, but the reading has to be built on a baseline of "what normal looks like". The coating itself produces regular interference reflections and the colour shifts with angle, which a newcomer easily reads as a defect.
Three checks separate them. Distribution: coating interference texture is even and continuous across the whole surface, while a real defect is isolated, random and irregular. Change with angle: interference colour shifts as a whole with the viewing angle, while the position and shape of a scratch or particle do not. Edge: a real scratch has a defined bright line or bright point boundary, while coating texture is a diffuse colour band. Build a baseline impression from good parts of the same batch, inspect under low ambient light, and close the spot to the 5 cm setting for confirmation; misjudgements drop noticeably.
Higher is not better. A polished mirror part under high illuminance produces strong glare and goes white across the surface, so the flaw is harder to see; too little illuminance leaves insufficient contrast and shallow scratches vanish. The right criterion is whether the brightness difference between defect and background is large enough, not a fixed lux number.
In practice, set it by material: mirror and cover-glass parts take a high-concentration configuration with a small spot and brightness reduced as needed; matt and rough surfaces can be brightened. The 10-99% stepless dimming of the 3318 exists for this - lower settings on low-reflectance samples to prevent glare - and its two colour temperatures, cold white at about 59000 lux and warm white at about 49000 lux, give another adjustment dimension. The push-pull zoom of the 3360 solves the same problem from the other direction: total flux stays much the same, the spot closes down and the intensity per unit area rises, reaching the 280,000 lux order at the 5 cm setting.
The angle is not one fixed value; it changes with the defect type, and this is the step most often skipped in surface inspection.
The "15°-30° oblique" often quoted in conversation refers to the recheck angle for scratch-type defects. The environment matters as much: work under dim light where possible, since ambient light directly eats contrast; move the head slowly and change the angle of incidence, which finds shallow scratches and pits more readily than sweeping at a single angle. Concluding after one angle is a common cause of missed defects at this step.
The difficulty with a curved surface is that the angle of incidence keeps changing, and holding the angle steady by hand is hard. Two approaches: follow the curve, keeping the light oblique to the arc so that reflection distortion reveals dents and bulges - a defect on an arc shows as a band of discontinuous light and dark under grazing illumination; or fix the lamp and move the workpiece, so the angle is reproducible and the criterion stays consistent.
A fixed station suits the 3318: the 500mm flexible arm adjusts height and angle freely, the 78mm head is positioned where it is needed and then left alone, and both hands stay on the sample; the 1.7kg base can be carried around the line. For handheld work use the 3360 at 340g with a 160mm body, held against the workpiece for long sweeps with one hand; for larger curved parts needing more coverage, the 3325G at 900g can go on a bench bracket.
Yes, provided there is no fan. An inspection lamp is itself a contamination risk in a clean area: a fan lifts dust, disturbs the airflow and blows light workpieces about, re-suspending particles that have just settled.
The 3318 uses a natural-convection aluminium housing with no fan, so it neither lifts dust nor disturbs airflow, which is the precondition for cleanroom use; the price is lower heat rejection than forced cooling, so it is not suited to running at full power for long periods inside an enclosed space. The 3360 series aluminium body is designed for cleanroom use, with a surface that is easy to wipe with a cleanroom cloth and cleaning agent, reducing the secondary contamination brought in with the lamp. Luminaires with a cooling fan belong at an ordinary inspection station, and should not be used in Class 100 or Class 1000 areas.
Two categories, matching two working rhythms. The zoom type is represented by the 3360: push or pull the body, and at 40 cm the spot corresponds to 5 cm, 10 cm or 50 cm, with no tool and no downtime. What changes is not the brightness setting but the spot area - total flux stays much the same, the spot closes down and the intensity per unit area rises. In use, open to the 50 cm setting and move across the workpiece to judge whether any bright point or bright line is present; when something suspicious appears, close to the 10 cm setting and recheck from two directions; close to the 5 cm setting for confirmation.
A fixed-focus version has better spot uniformity and suits standardised inspection at a fixed station, reducing operator-dependent variation. On the 3318 the 30cm working distance gives a 30cm diameter spot, and the head is positioned by a 500mm flexible arm, so once distance and angle are locked the criterion is comparable across a batch. Note that spot size depends heavily on working distance: for acceptance, measure at the working position rather than relying on the nominal figure in the sample.
It warms up, but the design already assumes continuous work. The 3360 uses an aluminium body for heat dissipation, so the heat from a 20W source running continuously is spread over the metal shell; it stays comfortable in the hand and the output does not fall off from heat build-up. The 3325G is a 36W high-power LED with a 180×200mm body that provides enough radiating area. Because the 3318 is naturally cooled with no fan, its temperature rise at full power over a long run is relatively higher - normal behaviour, and the price paid for not lifting dust.
Practical advice: a high setting running flat out is hotter than a low one, which is physics; put the lamp where air moves, and do not cover the head with cloth or paper to block the light, since trapped heat both shortens light-source life and creates a safety risk; where a station needs full power for long periods, choose a forced-cooling model that is not used in a clean area.
Yes, and they differ by form factor. The 3360 series ships as standard with one inspection torch, two rechargeable batteries, one AC charger and one aluminium carry case, and optional filters are available for background fluorescence or specific colour requirements. The 3325G takes a bench bracket to convert from handheld rounds to a fixed station. The core accessories for the 3318 are the 500mm flexible arm and the base, with an optional 500-600nm green filter.
Common accessories include a charging base, spare batteries, replacement lenses, mounting brackets, cantilever brackets and a carry case. Two points when specifying: fitting a green filter reduces illuminance, so measure at the working distance before deciding whether to move up a setting; and the bracket or cantilever should be kept away from strong vibration sources, or the spot will shake with the machine and long visual inspection becomes more tiring.
In the 3360 series, for example: a 5000mAh rechargeable lithium battery, 3-4 hours of continuous operation, 5-6 hours to charge, and two batteries in the standard package. The point of that arrangement is rotation - one in use, one on charge - so two shifts run without a gap; the metal button carries a charge indicator, so a glance before entering the cleanroom avoids a lamp dimming halfway through the work.
Where the work moves between machines away from mains power, the 3325G has a lithium battery version, the 3325GD, with the same specification as the mains-powered 3325G. On running while charging: continuing to run from the charger is workable at most stations, but a high setting over a long period while charging raises battery temperature, and two-battery rotation is the better pattern for continuous line work; for the supply arrangement of a specific model, state the workstation conditions when ordering and we confirm by model.
For long storage, keep a lithium battery at about 50% charge to avoid damage from deep discharge, and do not leave it by a window or next to an equipment exhaust in a hot workshop.
An LED source normally reaches 50,000 hours or more - more than a decade at 8 hours a day - and the practical limit is usually the battery and the lens rather than the source. The 3325G is specified at an LED life above 50000 hours, and the 3360 series likewise uses long-life LEDs, so no lamp replacement is needed for years.
Wear parts and repair: the lens and battery are wear items and can be purchased and replaced separately; a failed LED can be returned for repair, subject to the warranty terms. What not to do: cover the head with cloth to block the light, or run for long periods outside 0-40℃ (the stated operating range of the 3360 is 0-40℃); both shorten light-source life.
Oil and dust on the lens surface break up spot uniformity - dark patches, blackening or soft edges appear in the spot, which directly reduces detection. Clean with a cleanroom cloth and anhydrous ethanol, gently; do not use tissue, cotton waste or abrasive cleaners, which leave fine scratches in the coating.
A scratched lens produces stray light and a distorted spot: mildly, radial bright lines at the spot edge; badly, the judgement area loses uniformity, at which point the lens should be replaced rather than used as it is. The 3360 body surface is easy to wipe clean, and a quick wipe after each use is less trouble than cleaning a build-up later. The same applies to filters, which are consumables that get touched and marked; replace them when they cloud or scratch.
The portable models follow conventional industrial design for dust and splash protection, which handles workshop dust and a little moisture, but they are not suited to prolonged immersion, heavy oil spray or high-pressure washing - that class of environment belongs to protected-housing designs, and an ordinary inspection lamp should not be forced into it. For the IP rating of a specific model, go by the nameplate and the manual, and check against that marking at goods-in.
Two practical suggestions for an oily workshop: clean the body and lens after use, since oil film on the lens fogs the spot and lowers illuminance; and do not leave the lamp beside coolant or a cleaning tank, because oil mist that settles over time seeps into the buttons and seams. In a cleanroom version, mind how you wipe - a cleanroom cloth rather than a wet rag, to avoid leaving fibres.
Routine visual inspection has no mandatory metrology requirement, but two things are worth doing. Check the spot visually at intervals: switch on, shine it at white paper and look for blackening, dark patches or soft edges, which catches lens contamination and source decay early. Verify illuminance where a system audit requires it: measure at the standard working distance with a visible-light illuminance meter whose response follows the eye's V(λ) function; a UV radiometer cannot be used in its place.
For the meter, a combined UV-A and white-light or three-in-one model is more useful, since it verifies both visible inspection lamps and UV stations. In our range see the XP4000 UV radiometer (UV-A, white light and blue light in one). Record the measuring distance, the setting and the spot size together, or re-inspection of the same batch cannot be compared.
The optical principle and the inspection logic are the same - directional light at an oblique angle creates contrast, and the criterion rests on the brightness difference between defect and background - so there is no difference at the level of principle. The practical differences sit in three places: spot uniformity (beam shaping and diffusion differ, so the edge fall-off differs), condenser optics (spot shape at each zoom setting and consistency from centre to edge) and heat management and adjustment (temperature rise and output stability under long full-power runs).
On audits: the equipment can be supplied with spectral and illuminance parameter reports, which satisfy the appearance inspection station requirements of an internal quality system; where a customer needs a high-end benchmark comparison, samples can be provided for a side-by-side test, judging the same batch of known-defect coupons under each lamp once, and the result is more convincing than a parameter table. When selecting, treat "is the illumination method reproducible" and "is the criterion consistent across a batch" as the core measures - they affect the actual detection rate more than the origin of the brand.
The 18 questions compressed to one line each, for quick reference on site.
An ordinary lamp illuminates; an inspection lamp creates contrast: directional concentration, oblique illumination, glare control, even spot, no flicker.
Particles and micro-scratches: yellow-green 500-600nm. Colour and full appearance inspection: white 400-760nm. Photoresist yellow rooms: yellow 530-670nm.
No. Visible light finds geometric defects, and fluorescent materials need 365nm UV excitation; the two are complementary and not interchangeable.
Look at distribution and change with angle: coating texture is even across the surface and shifts with angle, while a real defect is isolated, random and fixed in position. Build a baseline from good parts.
No. A mirror part glares and goes white at high illuminance; set by material, and judge by the brightness difference between defect and background.
Flicker makes defects appear and disappear, causing misses; long inspection brings fatigue; and photographic records show banding that reads as a false defect.
A fan-free, passively cooled model can, since it lifts no dust and disturbs no airflow; fan models belong at ordinary stations and not in Class 100 or Class 1000 areas.
Particles grazing at close to 180°, fingerprints and oil film perpendicular at 90°, scratches perpendicular then 10°-30° off axis to recheck; low ambient light, slow movement of the head.
A zoom model changes 5/10/50cm spots by pushing the body; a fixed-focus model is more uniform and suits standardised inspection at one station.
The 3360 is 5000mAh, 3-4 hours of operation, 5-6 hours to charge, with two batteries in the standard package for rotation; the 3325GD is the lithium version.
An LED reaches 50,000 hours or more; lens and battery can be bought separately, a failed LED goes back for repair, subject to the warranty terms.
Wipe gently with a cleanroom cloth and anhydrous ethanol; a scratch causes stray light and a distorted spot, and a badly damaged lens should be replaced.
Visual inspection has no mandatory metrology; for a system audit, verify with a visible-light meter that follows the eye's V(λ) response at the standard distance, and record the setting and spot.
The portable models are dust and splash resistant and will not take immersion or heavy spray; the rating is as marked on the nameplate, and in an oily workshop clean the body and lens after use.
Follow the curve and adjust the angle of incidence, using reflection distortion to find dents and bulges; at a fixed station, lock the angle with a flexible arm and move the workpiece, not the lamp.
The principle is the same; the differences are spot uniformity, condenser optics and heat management. Parameter reports and side-by-side tests on the same coupons can be provided.
It warms up without becoming uncomfortable; a slightly higher rise on a fan-free model is normal, so do not cover the head and keep air moving.
Charging base, spare batteries, replacement lenses, mounting brackets, cantilever brackets, carry case; the 3360 ships with two batteries, a charger and an aluminium case.
Compressing the answers above into one selection rule: handheld sweeping, take the 3360G; walking the machine, take the 3325G; full inspection at a fixed station, take the 3318.
All three are in the surface inspection lamp section, where they can be compared by workpiece material and workstation form. If what has to be judged is fluorescent residue rather than a geometric defect, move to the UV route: MNT-365A for organic contamination rounds in a clean area, LUYOR-2130L for oxygen line degreasing acceptance, and LUYOR-3180 for fluorescent penetrant and magnetic particle testing; irradiance, white light and filter questions for NDT lamps are covered in 10 questions and answers on NDT lamps.
If the spectrum or the workstation form is uncertain, tell us the workpiece material, the defect type to be found, the inspection cycle time and the site lighting, and we will suggest a model and an illumination method for the actual conditions; sample coupons can also be sent for a comparative reading.
This article belongs to the FAQ section.
Other sections on the same product line: surface inspection lamps (low-angle side light that lifts particles, scratches and fingerprints out of the background) and industrial test instruments (the entry point for the four lines of UV inspection, speed measurement, stroboscopic observation and surface inspection).
Models mentioned in this article: LUYOR-3318, LUYOR-3325G, LUYOR-3360G. Range, accuracy and accessory lists for each model are given on the product pages.
Email: info@xinysun.com
Tel: 131-2093-6518
Address: 1525 Minqiang Road, Bldg 26, Songjiang, Shanghai
Copyright 2026 Xinysun Instrument (Shanghai) Co., Ltd. All rights reserved.
ICP:ICP-2026046447
