RTV Silicone Mold Surface Finishes Start on the Master

If you're specifying finish on a custom RTV silicone mold, you're specifying the *master*, because the rubber copies whatever you poured against. “High polish” in an email isn't a specification. SPI letters, VDI 3400, MoldTech / MT 11000 codes, and a measured Ra/Rz are languages toolrooms already speak for steel; on an RTV job they belong on the pattern (CNC, SLA/DLP, or machined metal), plus a note for the barrier coat and the jacket's inner face.

Pick one language per face, put it on the drawing, and say how you'll check the first casting. Platinum RTV, tin-cure, and the dump compound change how the rubber *releases*. They don't change the copy rule. Unmix food, optical, and “looks shiny” claims from the polish.

Specify the master, not a wish on the rubber

CNC or machined metal masters can carry SPI-style polish, a measured Ra on a sealing lip, or a cut grain. Shops still use SPI letters because both sides already speak them, even when no steel LSR cavity is involved.

SLA/DLP masters copy layer lines, sealer blush, and undercured inhibition into every dump. Full UV post-cure, wash, and a barrier coat are finish operations, not a courtesy. Polishing the mother mold after the pour doesn't rescue a layered print.

VDI 3400 and chemical textures are still valid callouts when the master is EDM'd or acid-etched, or when you're matching a European print. Don't write “VDI 18 or equivalent A-3” and assume the pattern shop will guess. They're produced differently and look different under silicone.

Ra / Rz (ISO 4287 and related practice) is useful on a lip that must release chocolate cleanly, or on a claim where a number can be audited. It's a poor way to specify a decorative leather grain on a soap mold. A grain can share an Ra with a dull polish and a completely different visual.

One face, one language. A gummy show face might be a named polish on the master. The same cavity's split might be a measured Ra so the keys shut. The grip wall of a bakeware mold might be a named grain. Mixing all three on one arrow is how finishers fight printers.

Soft platinum RTV likes a finish it can peel from without tearing a thin letter. A mirror master can increase stick on some low-durometer skins because you've removed the micro-vent of a duller face. A controlled dull often releases cleaner on consumer bakeware and gummy trays. That's a process observation, not a law — confirm on a coupon of the production RTV and the production dump.

Common mix-up: polishing the jacket instead of the master

A lot of RFQs treat “high polish silicone mold” as a job you do on the rubber after it's poured. In the shop that usually means the working face already copied layer lines, and you're shining the outside of the block. Finish the print or the CNC pattern (wash, UV, barrier coat, local sanding that doesn't kill letters) *before* the pour.

If someone quotes SPI A-1 on heated LSR steel with a cold-runner, that's a different quotation for a molded *part*, not this RTV mold.

A polish also doesn't buy food contact. FDA 21 CFR 177.2600 and LFGB are compound and extraction stories. A mirror master doesn't make a non-compliant dump food-grade. Cleanability helps hygiene in use; it isn't a substitute for the right polymer and a real test report on the article or a representative plaque. Infant/toy files (CPSIA) live on the article. Optical or “glass-like” resin only works if the resin is an optical grade, the master is truly A-range, the knit is managed, and yellowing is controlled. And a polish isn't a license to skip draft — silicone still tears on sharp, un-drafted, high-polish pockets in the master.

What to ask when you're buying

Send these before CAM starts:

• Which faces of the master get which language (SPI, VDI, Ra, named grain)?
• Is the master CNC, aluminum, or SLA, and what's the wash / UV / barrier-coat plan if it's printed?
• How will first castings be photographed, and is there a chip or plaque as witness?
• Do splits get a different finish from the show face so keys still shut?
• Food contact: extraction on the finished article, or only a grade letter?
• If we plate or metalize the master, is finish approved *after* that coat?

If the quote only says “mirror finish” and a price, you're not comparing the same job.

How to write it so the pattern shop can inspect

Put a finish note a CAM programmer and a QC inspector can both use: face name (show, split, logo pocket, core); standard and grade; method (CNC polish, hand finish, barrier coat, etch); witness (plaque, chip, or a designated first-casting photo under agreed lighting — a WeChat photo isn't a witness); direction of polish if it will read on a clear casting.

Stop polishing the master when the functional need is met. Past a functional A-range on a large chocolate bar you're buying hours the dump may not repay. Sealing lips and chocolate release want the lip specified, not jewelry on the whole pocket. Bakeware and grips want a specified grain more than “as polished.” Sticky low-durometer RTV often wants a tested dull master or a barrier coat before you chase A-1. A texture is a finish — treat it as a specified operation with draft and depth, and don't bury “add light texture” in a finish note.

Worked example. A six-cavity chocolate bar, 120 × 40 mm, 12 mm thick, food-contact platinum RTV with a resin jacket, needed a satin face so the bar released and a sharper logo pocket. The shop finished the CNC aluminum master to a B-range paper on the field, left the logo at a higher polish, and skipped a barrier coat because the master wasn't a print. First dumps proved the satin; a mirror trial coupon stuck on the low-durometer skin. The drawing now names both faces and a lighting setup for incoming photos.

When polish talk is the wrong RFQ

If the working face is already poured, polishing the jacket won't rescue layer lines — you recast after you finish the master. Food-contact isn't a diamond polish; if you needed FDA 21 CFR 177.2600 extraction, specify the test, not SPI A-1. Steel LSR show-face polish with a cold-runner is a molded-*part* quote. A tiny, shallow one-piece tray can live with a B-range master; a clear optical resin can fail even on A-range if the resin isn't an optical grade. Geometry and the dump override "shinier is better."

Related: texture on the master, how we make the mold, food-grade molds.

Finish callouts to send with the master

• Face-by-face language: SPI / VDI / Ra / named grain — one language per arrow
• Master process: CNC, aluminum, or SLA + wash / UV / barrier-coat plan
• Split finish vs show face (keys have to shut)
• Witness: chip, plaque, or lighting setup for first-casting photos
• Food-contact path if you meant hygiene, not gloss

Send master photos (or the STL) and which faces are show vs split. We'll mark finish, barrier coat, and whether a mirror will actually release your dump.

We would not recommend polishing the jacket instead of the master

We would not recommend chasing SPI A-1 on a low-durometer platinum skin as a default, or treating a diamond polish as an FDA 21 CFR 177.2600 path. We would not write “high polish silicone mold” as a job you do on the rubber after the pour. Finish the CNC or the washed, UV-post-cured, barrier-coated print first. We would not polish splits to the same grade as the show face if that rounds the keys.

RUUIPON Mold Structure Decision: one face, one language. Show face can be a named polish or grain. Split wants a slightly duller, intact shut. Soft 10A–20A or a wet chocolate dump? confirm release on a coupon before you pay for jewelry.

flowchart TD
  A[Which face?] --> B[Show / logo / split / lip]
  B --> C{Master type}
  C -->|CNC / metal| D[SPI letter, Ra on lips, or named grain]
  C -->|SLA/DLP| E[Wash + UV + barrier BEFORE pour]
  D --> F{Low-durometer stick on mirror?}
  E --> F
  F -->|yes| G[Step to B-range or tested dull; coupon the dump]
  F -->|no| H[Stop when the functional need is met]
  B --> I{Split?}
  I -->|yes| J[Duller intact face so keys shut]

Typical shop values

Caption these as typical shop values, not a named lot.

| Item | Typical shop value | | --- | --- | | Cavity Shore | 10A–20A platinum RTV, jacket behind it | | Mix | Platinum 1:1 or 10:1 by grade; vacuum degas until the foam collapses | | Platinum shrink after full cure | About 0.1–0.3% | | Show-face starting band on chocolate / gummy | Often SPI B-range satin; logo pocket may sit higher | | RTV skin wall | Often 8–15 mm behind a rigid jacket | | Loaded-tray edge sag if the jacket rocks | About 0.5–1.5 mm — reads as “uneven polish” on dumps | | Outer-row gummy vs center if the tray sags | Edge often 3–6% heavy |

FAQ

Is SPI A-1 required for food-grade RTV molds? Food-contact status comes from the elastomer and the extraction standard (FDA 21 CFR 177.2600, LFGB, and related files), not from a diamond polish. A sanitary, cleanable cavity helps. It doesn't convert a non-food grade into a food-grade article.

Can I call out Ra instead of SPI to be more scientific? You can, and you should on lips and splits that will be audited. Decorative or leather looks still need a VDI or MoldTech chip because two surfaces can share an Ra and not look alike. Use Ra where a number is the requirement; use a chip where the eye is the requirement.

Why did a high polish make the RTV harder to demold from the master? Some low-durometer platinum grades wet a mirror and lose the micro-air film a duller face provides. Confirm on a coupon. Adding draft, a tested barrier coat, or stepping from A-1 to a B-range finish often helps more than more release spray, which can contaminate food-contact lots.

Does polishing the jacket rescue a layered SLA master? No. The working face is the rubber that copied the master. Finish the print before the pour. Jacket polish affects the outside of the block, not the cavity.

Should splits be polished to the same grade as the show face? Usually not. Splits need a specified, often slightly duller, intact face that shuts on the keys. Over-polishing a split can round keys or make flash harder to see until it's a defect. Separate the arrows: show-face finish versus split condition.