How Shops Make Custom RTV Silicone Molds Step by Step
Custom RTV molds are made from a frozen master: box, keys, degassed pour, then a witness cast. LSR steel with a cold-runner is a different factory path.

If you're buying a custom RTV silicone mold, it isn't milled from a mood board. The shop finishes a freeze-ready master, builds a mold box, registers any split or core, mixes platinum- or tin-cure RTV A/B, vacuum-degases, pours, cures, pulls the master, fits a support jacket, and only then casts a witness article. Cutting steel for HCR compression or for LSR injection with a cold-runner is a different factory path. If you need a molded silicone *part*, that belongs on a separate RFQ, not as step six of this sequence.

If the CAD or the physical sample is still moving, you're in design. The sequence below assumes the pocket is frozen and the process is RTV mold-making.
Freeze the master before anyone orders a drum
The first shop document is the cavity: a STEP of the pocket, a 2D with the dimensions that matter, or a physical sample that will be copied. A pourer can't guess the split, and a mill can't guess shrink if a CNC master still has to be cut.
Before mix day, name platinum RTV or tin RTV (not "liquid silicone" as a blur), write shrink for *that* grade on the 2D including whether the CAD is already compensated, and note parting, pour direction, and peel direction. Shore A is a grade, not a color chip. Jacket / mother-mold material is agreed if the cavity will be soft (often Shore 10A-20A on food and gummy tools). A pretty render of a gummy mascot doesn't contain a shut face.
Make or accept the master
Surface finish of the master becomes the cavity finish. CNC aluminum or resin, SLA/DLP with full UV post-cure and wash, machined metal, or a finished customer sample are the usual sources. Layer lines, sealer blush, and undercured SLA inhibition show up as printed texture or sticky faces, so barrier coats and a platinum-inhibition checklist belong in this step rather than as a surprise after pour.
Tin-cure RTV is more forgiving of dirty masters and keeps shrinking. Platinum-cure RTV holds size better in storage and is the usual food-contact conversation, provided the master is clean of sulfur clay, tin residues, and some SLA resins. Choose the chemistry on purpose; neither choice turns this job into LSR steel.
If the program will later need thousands of molded silicone *parts*, design production draft on the article now even though the RTV mold can peel a zero-draft figurine. The RTV pour still doesn't qualify an HCR gasket or an LSR consumer part.
Box, split, keys, cores, then the pour
The mold box or flask sets pour height and wall thickness of the rubber. A warped plywood box prints as a tapered block. Registration keys (pips, bushes, or CNC bosses) are how a two- or three-piece mold finds itself every cycle; shops that rely on "the flash line will show me" watch the halves drift after a few hot dumps.
One-piece glove molds exist for shallow trays. Anything with a bail, a letter with negative draft, or a two-sided logo needs a split. Inner cores or mandrels for hollow chocolates and ice spheres need a pull path; if the core can't extract, you designed a cut mold. The support jacket is built with the box, not after the first slump. Plaster, resin, fiberglass, or a 3D-printed shell keeps the soft cavity from ballooning.
Shop order on the bench: release the master and the box; weigh RTV A/B to the grade's ratio, mix, and vacuum-degas; pour a thin stream, not a dumped cup; cure at the grade's conditions; demold the master; inspect the cavity; fit the jacket; then cast a witness article. Some small pours use a thin stream and vibration when a chamber is missing; that's a compromise against trapped air, not the production default when the pocket has fine detail.
A multi-piece RTV mold for a deep undercut is extra jackets and extra seams. It still isn't an LSR split tool. Don't print a master and then mill steel from the same file without a new shrink policy and call both "how the mold was made"; the numbers on the article won't match.
Common mix-up: treating the 3D print as the finished cavity
A lot of RFQs treat an SLA figurine on the desk as a production mold. In the shop that usually means the print still has layer lines, inhibition, and no jacket, so the first dump copies all three. The print can be the *master* after wash, full UV post-cure, and usually a barrier coat. Production still wants that master finished before pour.
What to ask when you're buying
- Is the master CNC, SLA/DLP, or a physical sample, and who owns the file after pour?
- Will you vacuum-degas the mixed A/B, and what's the cure schedule on the TDS you're using?
- One-piece, keyed split, or cored, and what is the jacket made of?
- Platinum or tin, and will the first witness be a dump of *my* compound (pectin, chocolate, PU)?
- If the first witness misses, is the next step a master refinish, a recast, or a new quote?
Worked example: 6-cavity fruit, about 1,000 dumps a month
A gummy line wanted a 6-cavity fruit, 22 mm tall, food-contact, about 1,000 dumps a month. The shop CNC-milled or fully post-cured an SLA master, built a keyed two-piece platinum RTV skin, and printed or laid up a rigid jacket so the sheet stays flat on the depositing line. First article was a dump-weight check on all six pockets, not a photo of pocket one. A later logo change is a recast against the same master if the master wasn't sanded "to clean it up."
When this sequence is the wrong path
If you actually need LIM flash and a platinum LSR *part* in the thousands, this pour sequence is a geometry study at best. Don't skip the steel tool. A tiny, shallow one-piece ice tray can skip a heavy mother mold. Geometry can override quantity: a 40 mm hollow chocolate with a stuck core is still a split-and-mandrel job at 50 pieces.
Related reading: what a custom RTV mold is, design before the pour, how long production takes, custom RTV molds.
Shop-sequence checklist to copy onto the PO
- Frozen pocket STEP or measured sample; shrink owner named
- Master finished (UV post-cure / wash / barrier if printed)
- Box square; keys, split, and core pull path drawn
- Jacket material agreed for the Shore you'll pour
- Mix ratio, vacuum degas, pour, cure per TDS
- Witness cast in the production dump; recast / refinish / release written down
Send the frozen STEP (or the physical master), the dump material, and photos of any undercut. We'll mark the split, jacket, and whether vacuum degas is non-negotiable on this pocket.
We would not recommend pouring on an unfinished print
We would not recommend pouring platinum RTV against an SLA master that has not been washed, fully UV post-cured, and barrier-coated. We would not skip vacuum degas on a pocket with serifs. We would not mill LSR steel from the same uncompensated file and call both "how the mold was made."
RUUIPON Mold Structure Decision: freeze the master first. Flat back? open-back. Deep undercuts that will not peel? keyed split or inner core. Soft 10A-20A or a wet dump? mother mold with the box, not after the first slump.
flowchart TD
A[Frozen pocket] --> B[Finish master]
B --> C[Box plus keys plus core path]
C --> D[Mix A/B, vacuum degas, pour]
D --> E[Cure, pull master, fit jacket]
E --> F[Witness dump of production compound]
F --> G{Miss?}
G -->|master finish| B
G -->|skin tear or wear| H[Recast against same master]
G -->|OK| I[Release]Typical shop values
Caption these as typical shop values, not a named lot.
| Item | Typical shop value | | --- | --- | | Mix | Platinum 1:1 or 10:1; tin condensation per TDS | | Cavity Shore | 10A-20A skin; jacket is the stiffness | | Platinum shrink after full cure | About 0.1-0.3% | | Vacuum | Degas mixed A/B until the foam collapses | | 6-cavity 22 mm fruit, about 1,000 dumps/month | Keyed two-piece platinum plus jacket |
FAQ
How is a custom RTV mold made if I only have a 3D print? The print can be the master after wash, full UV post-cure, and usually a barrier coat so platinum doesn't inhibit. Production still wants that master finished before pour. A printed figurine sitting on a desk isn't a cavity.
Do you mill a custom HCR or LSR tool the same way? No. Those are metal tools for molded *parts*, with their own shrink, heat, and runner hardware. This sequence is master, box, pour, and jacket. Ask for a separate quote if you need the part.
How are platinum RTV molds made without a vacuum chamber? Some small pours use a thin stream and vibration. That's a compromise against trapped air, not the shop default when the cavity has fine lettering or undercuts.
What is the actual last step of making the mold? A measured witness article against the frozen pocket, plus a written decision to recast, refinish the master, or release. A cavity photo isn't the last step.
Can a shop pour RTV as a shortcut to making an LSR mold? It can pour an RTV cavity to study shape. It can't skip a later LIM tool if production is molded LSR parts. Shrink, flash, and compound won't match, so treat the RTV step as geometry, then open a new RFQ for steel if you truly need parts.
Ready to order custom silicone molds?
Free quote for OEM/ODM projects. MOQ 20 · Prototype 3-7 days · FDA/LFGB available.
