← Blog

Niche sourcing

Reproduction Interior Vinyl: The Grain Is the Tooling

5 min read

You cannot order a period interior grain the way you order a colour, because the grain is not a property of the material — it is a piece of tooling. Coated fabric gets its surface from an embossing roll or a release paper, and the pattern that roll carries is a physical asset somebody paid for and holds. That settles most of the sourcing question. Taiwan has genuine, large-scale synthetic-leather capability, so a buyer's instinct to ask Taiwan for upholstery vinyl is not naive. It usually fails anyway, because a modern coating line's grain library is a library of contemporary automotive grains, and the period patterns this audience needs — the leather grains sold in the US trade under names like Madrid and Sierra — live as tooling inside the houses that already reproduce them. Here is what a competent material order specifies, and how to approve a sample whether you end up in Taiwan or in Oregon.

What are you actually specifying when you order interior vinyl?

A coated fabric is a stack, and every layer is a separate decision: a base cloth that supplies strength and structure, a coating that gives the material its body, a grain pressed into that coating, and a topcoat carrying colour, gloss and protection. The grain is the layer buyers misread — it is not a property of the chemistry, but a texture pressed in by an engraved embossing roll or by the release paper the coating is cast against. It is tooling: an asset somebody paid to have cut, and owns.

Around that stack sit the numbers a real order carries. Weight in ounces per linear yard is the shorthand for substance: Legendary Auto Interiors specifies 35 oz expanded vinyl for its seat upholstery and door panels, while general-purpose automotive vinyls in the Madrid and Sierra grains are sold at 28 oz on a 54-inch roll. Whether the material is expanded — carrying a foam layer between base and skin — changes how it takes a pleat and how it feels under a hand, and is the difference most often lost in translation. Then cold-crack temperature, published at −10°F for both those grains. Then UV stability, where the published figures are a trap of their own: the Sierra grain is quoted against a 1,000-hour xenotest and the Madrid grain against a 650-hour QUV exposure — different instruments, not comparable hours. Then flammability.

Flammability is not optional and not negotiable. FMVSS 302, at 49 CFR 571.302, applies to materials in the occupant compartment and names the components explicitly in S4.1 — seat cushions, seat backs, headlining, arm rests, all trim panels, head restraints, floor coverings, sun visors and convertible tops. S4.2 limits the requirement to any portion of a single or composite material within 13 mm of the occupant compartment air space. S4.3 sets the bar: the material shall not burn, nor transmit a flame front across its surface, at more than 102 mm per minute — and a specimen that stops burning within 60 seconds and 51 mm of the timing start is considered to pass.

What a reproduction interior material order has to say, and how to check it on the sample
What to specifyWhy it matters on a restoration partWhat a vague order gets youHow to check it on the sample
Grain pattern (by name or by physical master)It is embossing tooling, not a finish option; the wrong grain is visible from three metresThe mill's nearest house grainLay it on a known original under raking light; compare pebble size and depth, not colour
Gloss levelPeriod interiors read matte or satin; a modern high-gloss skin looks like a boat seatWhatever the standard topcoat givesSide by side with the original at a shallow angle; photograph both together
Colour, under two illuminantsMetamerism — two greens that match in daylight and diverge under shop lightA digital colour code that means nothing across chemistriesOutdoors and under your shop lighting, against a piece you are not replacing
Base cloth and weaveSets stretch, sew-through behaviour and how the panel holds a shapeCheapest available knitCut a corner and look at the reverse; pull it on the bias
Expanded layer, and its thicknessDecides whether a pleat stands up or collapses; the commonest silent substitutionUnexpanded material at the same weightSection it; press it with a thumbnail and watch the recovery
Weight in oz per linear yardThe trade's shorthand for substance and durabilityA lighter roll that still meets the colour specWeigh a measured piece; do not accept a spec sheet alone
Cold-crack temperatureA car stored cold and a door slammed in FebruaryA summer-only materialAsk for the test report; figures such as −10°F are commonly published on trade vinyls
UV and heat ageingA parcel shelf and a seat top live under glassFade and chalking in two seasonsHours and the test method in writing — a 1,000-hour xenotest and a 650-hour QUV are not the same number
FMVSS 302 burn rateLegally required of occupant-compartment materialsAn untested industrial vinylA 102 mm/min test report on that construction; the 13 mm composite rule applies
Roll width and usable yieldNesting a pattern on a 54-inch roll versus a 60-inch one changes yardage per carA quote that ignores wasteNest your largest panel on paper before you buy

Why is Taiwan a real coated-fabric country and still usually the wrong place for this?

The capability is not in question. San Fang Chemical Industry was founded in Kaohsiung in 1973 and, per the Taiwan Textile Federation's supplier directory, was the first company in Taiwan to manufacture wet-process PU synthetic leather; its own site describes over fifty years in the business, factories in Taiwan, China, Vietnam and Indonesia, and automotive among the industries it serves. Nan Ya Plastics publishes an automotive synthetic-leather line covering PVC and PU constructions and microfibre, quotes ISO/TS and IATF 16949 among its certifications, and lists seats, instrument panels, armrests, door panels and steering-wheel covers among its applications. Both are large listed manufacturers, cited as public evidence of what the island can do — not as sourcing leads.

The problem is that every one of those lines is tuned to a contemporary automotive programme. Its grain library is what its OEM customers have specified in the last two decades. Nobody is holding an embossing roll for a 1960s American pebble grain, and nobody will cut one for you: roll tooling is amortised over programme volumes, and a few hundred yards of restoration yardage does not begin to pay for it. Ask, by all means — but ask "what is already in your library that is close?", not "can you make this grain?", because the honest answer to the second question is no.

The duty picture does not help either. Imported yardage lands in the textile chapters rather than the Free seat-parts headings: PVC-coated fabric under 5903.10 ranges from Free to 14.1 percent by construction and plastics content by weight, and PU-coated fabric under 5903.20 tops out at 8 percent, both read live from the USITC tariff database on 3 August 2026. Which breakout applies to a given roll is a laboratory determination — ask your broker before you order.

How do you approve a grain and a colour when you are not in the room?

You do not. That is the schedule risk nobody prices in. Grain depth, pebble size, gloss and colour cannot be judged from a photograph, a scan or a digital colour code, because the thing you are matching is a physical original whose own surface has aged. The only method that works is physical samples in your hand, against a known-correct reference, viewed in daylight and again under the lighting of the shop that will fit the part.

That means a sample loop across an ocean, measured in weeks each way plus the mill's queue, and it usually means more than one loop. Build that into the plan before you commit to a car with a customer waiting for it.

  1. Send a physical reference — a piece of the original material, not a photograph, and say which face is the show face
  2. Ask for the grain by the mill's own library reference once you have seen it, so a repeat order reproduces it
  3. Judge grain and gloss first, at a shallow angle beside the original; colour is the last check, not the first
  4. Check colour outdoors and under shop light before you approve — metamerism is what makes a repaired seat look wrong
  5. Run a cold-crack and a flame test on the approved sample, and keep a retained piece as the reference for incoming inspection
  6. Nest your largest panel on the actual roll width before agreeing a yardage price

When is a US material house the only right answer?

Whenever the grain you need is not already in somebody's library, and always when the original was cloth. SMS Auto Fabrics in Canby, Oregon says it weaves over 1,000 different patterns of cloth, makes over 1,000 styles of door panels and stocks thousands of vinyls, covering classic American and selected foreign cars from the 1930s through the 1990s. That is an archive of tooling and patterns accumulated over decades. It cannot be commissioned, only bought from.

For an exact original cloth — a broadcloth, a bedford cord, a wool or a mohair woven to a period pattern — there is no overseas path at restoration volume at all. Reproducing a woven cloth means warping a loom to a pattern, and the houses that already did that work are the market. The same is true of most period vinyl grains. Taiwan becomes interesting only when you are launching in programme quantities and can either live inside an existing grain library or fund tooling as a capital item.

Our own position, plainly: TaiSource's structured catalogue is automotive hard parts. A materials project is manual legwork for us — visiting coating lines, walking their grain libraries, carrying samples — and we will tell you before you pay if a US material house serves you better. If you are launching a line and need a coating partner at volume, describe the material on our request page: vehicle, years, what the original was, weight, expanded or not, and the reference you can send. Taiwan Supplier Hub returns a shortlist of at least three verified manufacturers that have confirmed they want the project within 14 days, US$99 at launch, refunded if we can't deliver.

Have a real part to source? Describe it in plain English.

At least 3 verified, willing suppliers within 14 days — or your money back. Launch price: US$99 per request.

More from the blog