The second shipment arrives. The goods-in inspector cuts the crate strap, lifts the first enclosure out, wipes the packing dust off the door and stands it beside a unit from the first batch that has been sitting in the assembly bay since the spring. Same drawing, same RAL code, same factory, and the door is a shade greener and visibly flatter. Neither part is defective on its own: each one would pass inspection alone. What is missing is a powder coating colour match specification: colour, gloss and texture, written before the first batch and remembered on the second.
What is a powder coating colour match?
A powder coating colour match is a three-part specification (colour, gloss and texture) agreed before coating starts, then re-checked on every later batch against a signed limit sample. A code number on its own is not a specification, because it fixes none of the three: colour is the CIELAB position, gloss is reflectance at a stated angle, and texture is the finish family.
Why powder coating batches drift: six causes
A powder line does not repeat itself. It is a chain of variables, each with a range. Batch one lands somewhere inside those ranges, batch two somewhere else; the shifts accumulate until the eye sees them.
| Cause | What it does | Where it shows up |
|---|---|---|
| Powder lot change | Pigment grind and resin differ slightly between lots, even within one supplier's tolerance | Whole surface shifts together; check this first |
| Film thickness distribution | Thicker film reads darker; thinner reads lighter and lets pretreatment show through | Flats differ; edges differ from flats |
| Cure schedule and oven load | Underbake leaves the film matte; overbake darkens, yellows or shifts gloss | Light versus full oven; thick versus thin sections on one rack |
| Substrate and pretreatment | Steel, galvanised, aluminium and zinc-rich primer change what returns to the eye | Mixed substrates; welds read differently |
| Reclaim powder ratio | Reclaimed fines have a different particle size distribution and can carry contamination | Gradual drift, or a step change |
| Application parameters and texture | Voltage, flow, distance and pass count change film build and how a texture develops | Texture changes character; recesses thin out |
None of these is a defect: all six can sit inside normal practice and still produce a visible difference. Craters, pinholes and contamination show up on one part and get rejected on their own, and they are covered in the powder coating defect troubleshooting guide. What follows here is the harder kind: two parts that each pass inspection and fail only when they stand next to each other.
Specifying colour: the reference is not the spec
"RAL 7035" tells the powder supplier which target to aim at. It does not say how far the finished part may sit from that target, under which light, or on which substrate, and that gap is where goods-in arguments start. RAL is a commercial colour reference collection (RAL CLASSIC, DESIGN and EFFECT) administered by the German RAL gGmbH, not an ISO, EN or national standard.
Three consequences follow. Two suppliers making the same code do not make the same colour, because each formulates to the reference as its pigments and resin allow, so name the powder supplier and product code on the drawing. The same code at a different gloss is a different object. And the substrate is part of the colour: the four common ones in this table do not return the same thing to the eye (sheet metal material selection guide).
The code does not travel through the process: film thickness and cure move the result away from a reference panel. A usable powder coating colour specification has four lines:
- Reference and powder — the RAL code, plus the powder supplier and product code.
- Delta E limit with its conditions — colour space and formula, illuminant, observer, instrument, aperture, specular mode.
- Gloss target and tolerance at a stated angle — 20°, 60° or 85°, in gloss units.
- Substrate of the limit sample — plus the film thickness window and cure schedule it was made at.
Metamerism is the one that catches people out. Two panels can match under one light and separate under another when the formulations reach the same colour by different pigment routes. Judge the sample under the light the product will be seen in.
Gloss: why a powder coating colour match needs a second number
Gloss is specular reflectance, how much light returns in a mirror direction, measured in gloss units at a stated geometry by the methods in ISO 2813 or ASTM D523. Both define three geometries: 60° is the universal angle, 85° is used when the 60° reading falls below about 10 GU, and 20° when the 60° reading exceeds about 70 GU. One surface therefore gives three numbers. Quote a gloss number without its angle and nobody can check it.
Gloss and colour are independent properties, measured separately. A colour reading says nothing about sheen, and a RAL code says nothing about whether a part is matte, satin or gloss. It has happened to us more than once: a batch measures inside the agreed delta E and gets rejected anyway, because the sheen has dropped and the parts look dull beside the earlier ones. The instrument was right. So was the customer.
Texture is the third axis and the one most often missing from drawings. Finish families — smooth, satin, matte, fine or coarse texture, hammer, sand, wrinkle, metallic, clear over base — change appearance at identical colour and gloss, because they change how light scatters. Textures also hide substrate variation and handling marks better than smooth gloss, which is why we push them for large flat doors. The trade-off is cleaning: a coarse texture holds dirt, and an enclosure that gets wiped down daily is easier to live with in a smooth finish.
Neither number captures a fourth property, and it is why a surface can measure correct and look wrong: distinctness of image, or orange peel where it is coarse enough to see. We zone a product into visible and non-visible areas and apply limits per zone, as set out in the cosmetic surface class guide. Agree the gloss band with the applicator at a stated angle; matte and textured finishes hold across batches more easily than high gloss.
Delta E in powder coating: the five conditions to state
Delta E is one number: the distance between two colours in a three-dimensional space. Buyers regularly ask us to turn it into a grade, and we decline, because the number only means something once the illuminant, observer, geometry and instrument are fixed too. No single pass/fail value survives that, which is why we will not print one.
The space in common use is CIELAB — L\ lightness, a\ green-red, b\* blue-yellow. A measurement gives one point, the reference another, and the distance by a formula of the kind in ISO 11664 or ASTM D2244 is the delta E. Its usefulness depends on the conditions:
Five conditions a delta E figure has to state:
- Illuminant — daylight of a stated temperature, a specified fluorescent type, or incandescent. A metameric pair can be close under one, far under another.
- Observer — the standard observer function used in the calculation. Rarely disputed; still write it down.
- Geometry — specular included or excluded, plus the aperture. Small apertures read the peaks of a texture, large ones average the valleys, and the two do not give the same answer.
- Instrument and calibration — two makes need not agree on a textured or metallic finish. Compare batches on one instrument, or cross-check both against one tile.
- Sample condition — clean, dry, fully cured, and read on the same type of area: a flat, not an edge, curve or weld.
A delta E without those five is a number with no address; the disputes we see almost always involve a figure produced under different conditions on the two sides. A trained eye on large adjacent flats also flags what an instrument calls small, because it compares two large areas side by side — a stricter test than a 25 mm aperture on two panels.
Ask instead for a limit agreed with the applicator — for a specific powder on a substrate, on a stated instrument and illuminant — recorded on the drawing, with both batches measured the same way. A tight limit is realistic on a smooth single-colour powder on one substrate. On a coarse texture sprayed over both steel and galvanised, the same number turns into a weekly argument.
Film thickness: the quiet variable
Film thickness shifts a colour without anyone noticing. It is invisible on the drawing and on the part, and obvious the moment two parts stand together.
A typical working window for single-coat polyester on steel is around 60–120 µm, agreed per powder and project. Treat it as a window. The applicator is not aiming at 80 µm; he is aiming at a spread across the whole part, and that spread is what you see. Measurement is by gauge on the cured film: ISO 2178 or ASTM D7091 on steel by magnetic induction, ISO 2360 or ASTM D7091 on aluminium and other non-magnetic conductive substrates by eddy current, with the thickness method itself named per ISO 2808. Confirm the window against the powder supplier's technical data sheet.
Under-film reads lighter, and at the low end the pretreatment shows through: a grey over dark zinc primer is not the grey over a light phosphate. It stays coarser in flow, so gloss drops, and corrosion performance goes with it — a worse failure than appearance (corrosion resistance guide, salt spray testing guide). Over-film pushes colour darker and, on light colours, yellow or muddy; very high build drops gloss and brings brittleness at bends.
Geometry decides where the thickness goes. Faraday cage areas, meaning inside corners, deep recesses and boxed sections, receive less powder because field lines will not enter, and they come out thin and lighter.
Edges attract more powder, but surface tension draws the film thin during flow-out, so a sharp edge is the classic thin-coverage site (hang points and masking guide). A part can measure in range on the flat and still show edges and recesses that differ from it. Thickness, cure and gloss cannot be specified independently: a powder reaching its gloss mid-window may not reach it at the window's edges.
Cure: overbake and underbake
Cure is where powder melts, flows, levels and crosslinks, set by part metal temperature and time against the supplier's schedule. The oven air readout is not the part metal temperature; on a loaded line the two differ substantially.
Underbake means the film never reached its condition. It stays under-levelled, so gloss falls and the flow pattern stays coarse, and colour reads lighter. It has also not reached its designed resistance: adhesion at risk, impact and solvent resistance down.
Overbake means more temperature or time than the schedule. Light colours yellow or shift warm. Gloss rises on some chemistries and falls on others, so more bake is not a way to buy gloss. The film is more brittle, and reduced UV stability means the part starts service already partly aged.
Both can occur in one rack: a light 1 mm door reaches temperature early and sits there while a heavy 3 mm frame arrives late. That is a loading and line-speed problem. It is also why we ask for production parts at first article instead of coupons: a 100 × 50 mm coupon reaches temperature in a minute and says nothing about a 3 mm frame hanging next to it.
Reclaim powder and lot control
Powder that misses the part is collected and fed back into the hopper blended with virgin powder. That is reclaim. At production overspray rates it is unavoidable, and it is also a source of drift.
The mechanism is mostly particle size. Reclaim has been through the gun, the booth airflow and the recovery system, and the fine fraction fluidises and charges differently. A high ratio therefore changes film build at a given gun setting, and shifts colour and gloss with it. It also carries contamination: traces of a previous colour, fibres, dust, moisture. Ask for:
- Reclaim ratio — whether reclaim is permitted at all, and at what maximum proportion of virgin. For a brand colour or a metallic, the answer is often virgin only, and it is cheaper to agree that as a cost line up front than to discover it at goods-in.
- Powder lot number — on the job card and on the packing list. Six months later it is the first thing you compare.
- Lot reservation — reserve enough of one lot for a multi-batch programme. Store it sealed, dry and out of the sun; powder ages.
- Cross-batch panels — keep a panel from each lot next to the panel from the lot before it.
- Colour change discipline — booth, guns, hopper and recovery cleaned between colours. For a light colour running after a dark one, that cleaning is the difference between a clean job and specks.
We got this one wrong the first time, and it is now the first thing we check when gloss drifts. Repeat batches of a textured grey were going matte over several weeks; we assumed reclaim, capped the ratio, and the drift continued. The cause was a partly blocked fluidising plate: less powder out at a fixed gun setting, lower build on the flats, lower gloss. We now check the plate before we touch the reclaim ratio.
Assembling parts from different batches
Even with a sound specification, projects end up with two coating batches in one assembly: a repeat order eighteen months later, a spare door, a second cabinet in a row. Sequencing and tagging decide whether that matters, and this is where cross-batch consistency is won or lost.
Sequence the build, not just the paint: parts for one cabinet coated in one run and kept together through assembly. It fails routinely, because parts are kitted by part number, so doors from one lot meet bodies from another at the station. Tag by lot number, not by date: a lot number tells you which powder was in the hopper and which oven log to pull, and it is what the packing list has to carry (export packing and crating guide).
Avoid putting two lots on one visible face. Two adjacent panels from different lots get compared every day for the life of the product; the same two panels on cabinets in two rooms never get compared at all.
Rework economics: repainting one panel to match a two-year-old cabinet is a poor bet, since the original has weathered and the new part has not. Repainting the whole visible face is often cheaper, because a single-panel attempt usually fails and you pay for strip and recoat twice. We blast to a defined profile rather than burning off. Burn-off leaves ash in weld seams and along hemmed edges, and under a new film that shows up as adhesion loss at exactly the places you cannot rework. Keep one limit sample where parts are actually compared, at goods-in or on the assembly bench, not only in the paint file.
Weathering: colour change over years
Every organic coating changes outdoors, and the resin family sets most of the answer. UV attacks the resin at the surface, the binder degrades and releases pigment, and the surface dulls and then develops a powdery residue, known as chalking and assessed by methods such as ASTM D4214. Colour then reads lighter as the chalk builds. The three families differ like this:
| Resin family | Colour and gloss retention | Cost | Where we use it |
|---|---|---|---|
| Polyester | Holds colour and gloss reasonably; chalks eventually | Baseline | The default on cost in temperate climates |
| Super-durable polyester | Better gloss retention and slower colour change under UV | Higher per kilogram | Strong sun, or a brand colour that has to stay recognisable |
| PVDF | Holds colour and gloss longest of the three | Highest | Where the buyer has to hold an appearance retention commitment |
PVDF is a factory-applied fluoropolymer system with its own application and cure regime, so it is not a drop-in substitution on a polyester line.
We will not turn that into years. Weathering depends on UV dose: latitude, orientation, temperature, salt load and how often the surface gets washed. The same powder on the north wall and the south wall of one building is not in the same test.
What you can get instead is a comparison: xenon arc under ISO 16474-2 or ASTM G155 ranks two coatings run together, which is useful for screening and is not a calendar-life prediction. We can run it or arrange it, and report the measured colour and gloss change on the exposed panels. We do not certify the result, and the last section of this article sets out exactly what we will and will not put in writing. Geometry matters more than most specification arguments: spend on resin grade and on shading (outdoor sunshade and hood design guide), not on a tighter initial tolerance.
Limit samples and sign-off
The limit sample is what settles disputes. Codes, numbers, tolerances and angles all exist to describe it.
Approve a sample carrying all three axes: colour, gloss and texture, on the production substrate, at production thickness, from the production cure. Insist on a production part, or a coupon cut from one. A sprayed card carries none of the thickness distribution or edge thinning that makes two batches look different side by side.
How many: three. That is a working minimum: one master with us, one with the customer, one where parts are compared.
Store them sealed, opaque, dry, at room temperature and away from the line; label with date, powder lot, substrate, gloss reading and instrument. Re-issue on lot, supplier, substrate, thickness, cure or texture change, and when the sample has aged past what the line now produces.
In use, put the incoming part beside it under the lighting the product will be seen in, alongside an instrument reading taken the documented way; if the sample and the instrument disagree, the sample wins. Our records follow the same discipline as the rest of our inspection work, with limits issued per zone as described in the cosmetic surface class guide.
Paste-ready specification block
COLOUR & APPEARANCE — fill in and attach to the drawing
1. COLOUR REFERENCE
Reference system and code: RAL ______ (or supplier product code: ______)
Powder supplier / product code: ______ / ______
Metallic / effect: none | metallic | other: ______
2. COLOUR TOLERANCE
Delta E limit: ______ (agree with applicator; not copied from another job)
Colour space / formula: CIELAB, per ISO 11664 / ASTM D2244: ______
Illuminant: ______ Standard observer: ______
Instrument & geometry: ______ Aperture: ______
Specular: included | excluded
3. GLOSS
Target gloss: ______ GU at ______° (20 / 60 / 85)
Gloss tolerance: ± ______ GU (agree with applicator)
Method: ISO 2813 / ASTM D523
4. TEXTURE / FINISH FAMILY
Family: smooth | satin | matte | fine texture | coarse texture
| hammer | sand | wrinkle | clear over base
Supplier texture reference: ______
5. FILM THICKNESS
Target and window: ______ µm (60–120 µm typical single-coat polyester)
Measurement method: ISO 2178 (steel) / ISO 2360 (aluminium) / ISO 2808
Minimum on Faraday-cage areas: ______ µm Minimum on edges: ______ µm
6. SUBSTRATE & CURE
Substrate / pretreatment: ______ / ______
Cure schedule (part metal temp / time): ______ °C / ______ min
Reclaim powder: not permitted | permitted, max ______ % reclaim
7. LIMIT SAMPLE
Sample ID and date: ______
Issued by / held by: ______ / ______
Re-issue triggers: lot, supplier, substrate, thickness, cure or texture change
8. ZONING
Zones per drawing; limits apply per zone.
9. WEATHERING (outdoor product only)
Resin family: polyester | super-durable polyester | PVDF
Exposure comparison (if used): ISO 16474-2 / ASTM G155 — comparative, not a life predictionFrequently asked questions
What is a realistic powder coating colour match tolerance?
There is no universal number, and any single delta E value quoted without its conditions should be treated with suspicion. What is achievable depends on the powder system, the finish family, the substrate and the measurement geometry. Agree a limit with the applicator for that specific combination, record it on the drawing with illuminant, observer, instrument and aperture, and verify it against a signed limit sample. A tight limit is realistic on a smooth single-colour powder on one substrate; the same number on a coarse texture across two substrates is another matter.
Can two batches made with the same RAL number be guaranteed to match?
No. A RAL code is a reference, not a specification: it names a target colour and controls neither the powder supplier's formulation nor film thickness, gloss, texture, substrate or cure — every one of which moves the finished appearance. Two batches from different powder lots, or the same powder at different film build, can both be sold legitimately under one code and still look different side by side. Cross-batch consistency comes from a specification plus a limit sample, not from a code number.
Why does the gloss change when the colour measurement says the parts match?
Because gloss and colour are separate properties measured separately. A spectrophotometer reports colour position in a defined space; gloss is specular reflectance at a stated angle per ISO 2813 or ASTM D523. A film can sit inside its colour tolerance and still be duller or shinier than the reference, usually through film thickness, cure or reclaim ratio. This is the commonest rejection we see where the instrument is right and the customer is also right, and it is why gloss needs its own line on the specification with its own tolerance.
How much reclaim powder is acceptable if appearance matters?
It depends on the finish and on how critical the appearance is. Reclaim changes the particle size distribution and can carry contamination, and its effect on appearance comes mainly through film thickness and gloss. For a brand colour, a metallic or large flat visible surfaces the sensible answer is often virgin powder only, agreed as a cost line up front. For general industrial finishes a capped reclaim ratio with lot records is normally workable. What matters is that the ratio is stated on the order rather than left to the operator.
What should a limit sample look like and how many do we need?
It should be a production part, or a coupon cut from one, carrying the real substrate, film thickness distribution, cure schedule and texture. Not a sprayed card. Three is a working minimum: one master retained by the applicator, one with the customer, one at the location where incoming parts are actually compared. Store them sealed, dark, dry and labelled with date, powder lot, substrate, gloss reading and instrument. Re-issue on any lot, supplier, substrate, thickness, cure or texture change, and date every sample so an aged one is not mistaken for the master.
Can you match a panel to an enclosure that has been outdoors for two years?
Sometimes, but it is a poor bet. The existing surface has weathered: gloss loss, chalking and usually a colour shift. Fresh powder has not. Matching it means reproducing an aged appearance with a new coating, which takes a blended sample and a skilled eye, and it will still drift as the new panel weathers. Repainting the whole visible face is often cheaper than a second attempt at one panel, because the first attempt frequently fails and you pay for strip and recoat twice.
Does a tighter colour tolerance make the coating last longer outdoors?
No. Colour tolerance controls how close the delivered part is to the reference on the day it is made; weathering is governed by the resin family, the UV dose and the site. A very tightly matched standard-polyester part will still chalk and lose gloss outdoors, while a normally tolerated super-durable polyester or PVDF part will retain appearance better. If long-term appearance matters, spend on the resin grade and on shading rather than on tightening the initial tolerance.
Do you certify outdoor weathering life or warranty the coating system?
No. We apply and finish the coating, and we set, log and hold the process variables in this article — powder lot, film thickness, cure profile, gloss and colour measurement, limit samples. We can run or arrange comparative exposure testing under ISO 16474-2 or ASTM G155 and report measured colour and gloss change on the exposed panels. We do not issue coating-system warranties or weathering certifications for the end product, because service life depends on site, orientation and service conditions we do not control. Better to say that at quotation stage than argue about it later.
The short version
Cross-batch consistency in powder coating comes from three specifications, not one: colour, gloss and texture. A RAL code names a target and controls nothing else. Put all three on the drawing with tolerances, measure both batches the same way, and hold a limit sample taken from a production part. Everything else in this article is detail on how to do those four things.
Where we fit
We are a sheet metal OEM/ODM factory in Shenzhen. We cut, form, weld, prepare and coat enclosures and cabinets, and the finishing line is in our own building, so the people setting gun parameters, oven profile and reclaim policy are the ones who answered the drawing.
So what we sell is a documented process rather than a certificate of appearance. Specification is agreed at drawing review. First-article parts run on the production process, not a lab set-up. Colour and gloss are measured under stated conditions, cross-batch panels are retained, lot numbers recorded and the cure profile logged (quality control and inspection). We do not issue coating-system warranties or weathering certifications for the end product: we can run comparative exposure, report measured change and recommend a resin family, but how long a coating lasts depends on your site conditions.
Related reading: the finishing menu and what powder does and does not do, in the surface finishing guide and the powder coating versus liquid painting comparison; aluminium's alternative route in the anodizing guide; and the rest of the cluster in the materials and finishing hub.
Written by the Sanjun finishing team, Shenzhen. Last reviewed 11 October 2026.