RTV Mold Shrinkage Factors: Master, Rubber, Casting

If you're buying a custom RTV silicone mold, shrinkage isn't one catalog percent off a TDS. You're stacking three offsets: the master (if it's still post-curing resin), the mold rubber (especially tin-cure), and whatever you dump or pour into the cavity — pectin, chocolate, wax, polyurethane. That last layer is usually the big one.

Here's the part that usually bites people: an LSR heat-cure factor from a molded-*part* quote will oversize your master. Answer chemistry first (platinum addition-cure or tin condensation-cure), then pour a coupon in the production RTV lot, at the production mix, degas, and cure, and measure the *casting* in the production dump at a stated temperature. Until that coupon exists, catalog midpoints are just a planning band.

Three layers, three numbers

Master. CNC aluminum, machined resin, or a fully washed and UV-post-cured SLA/DLP pattern is usually treated as frozen once finishing is done. If the print is still outgassing, or the sealer is too thick in the lettering, it isn't frozen. Keep that master if you want a recast next season; sanding it “to clean it up” changes every later cavity.

Mold rubber. Platinum RTV used as a flexible mold typically shows very little linear change after a complete room-temperature or mild-oven cure, which is why shops cut masters near nominal plus whatever the *cast* article will shrink. Tin-cure keeps moving as byproducts leave, so treat it as time-dependent. Fine for a short run or a dirty master; not a library dimension you reopen in six months.

Cast article. Chocolate, gummy, wax, and polyurethane each have their own freeze or cure shrink. Vacuum-cast stacks accumulate shrink from each layer: master, RTV cavity, then the resin. Budget each layer. A process-total percent isn't a property of the gum.

CTE is reversible motion while the dump is still hot. Shrink is what you measure cold after it has set. Don't stack CTE×ΔT on top of a shrink factor as a second cut allowance on the master — the coupon already cooled.

If a quotation talks about heated cavities, cold-runners, and a 2–3% class LSR factor, that's a different RFQ for a molded silicone *part*, not this RTV mold.

Platinum, tin, filler, and wait time

Once you've named the family, shop variables still move the offset.

A soft Shore 10A–20A skin against the master doesn't move the same way as a filled, higher-tear jacket rubber if you poured them from different lots. Food-contact work under FDA 21 CFR 177.2600 or LFGB constrains which grades you may use; it doesn't freeze shrink.

Pigment in the *mold* rubber is uncommon, but color in the *cast* article (gummy, resin, soap) can shift fill and apparent size. Match the production dump, not a water test.

Wait time is the tin-cure problem. You can freeze a short-run process: pour the mold, wait a defined number of days, measure a casting, then dump. You can't treat that measurement as a library dimension months later. For fits that must recast next season, use platinum RTV and keep the master.

If you post-cure the mold, orientation in the oven matters. A thick jacket wall and a 1 mm lip at a logo don't freeze alike. An isotropic CAD scale on the master is a first cut, not a truth on mixed sections.

Common mix-up: putting an LSR percent on an RTV master

A lot of RFQs treat “silicone shrink ~3%” as a cavity number for any silicone job. In the shop that usually means someone pasted a heat-cured LSR planning band onto a platinum glove mold. Platinum RTV after a full cure is near-zero; the dump's shrink is what you compensate on the master. If you apply the LSR number, the pin, the logo, or the fit goes the wrong way.

ISO 3302-1 class (M1/M2/M3) tells you how tight a *rubber article* on a drawing may be. It doesn't tell the pattern shop the shrink percent of a gummy or a chocolate, and it isn't another word for “precision.” ASTM D412 / D624 characterize tear of the mold rubber; they don't replace a shrink coupon of the casting.

What to ask when you're buying

Send these with the RFQ:

• Platinum or tin, and at what age will you measure the first casting?
• What's the dump or pour compound (pectin, chocolate, PU, wax), and do you already have a coupon of *that*?
• Measure at what temperature, after what wait? 23 °C after a named wait is the usual lab state.
• Is the master getting compensation for the cast article only, or did someone also add an LSR factor?
• Inner core or jacket constraint — which dimensions are locked, which are free?
• If we recast next season, is the same master still in custody, unsanded?

If two quotes use different chemistries or different wait times, you're not comparing the same factor.

Geometry, the jacket, and a worked example

Thick sections in the *casting* hold heat longer and often show a larger cold offset than thin membranes. Ribs, undercuts, and an inner core constrain the dump while it cools. A hollow chocolate locked on a mandrel doesn't shrink the way a free gummy does; the ID may be set by the core while the free face pulls.

The support jacket (plaster, resin, fiberglass, or a printed shell) is how a soft cavity holds size during the pour. A slumping glove without a jacket prints a different shrink every dump, which looks like a “bad factor” on the CMM and is actually a missing mother mold.

Multi-cavity RTV tools can show cavity-to-cavity size drift if pour head, degas, or jacket registration differs. That's a process problem, not a reason to invent a second factor.

Worked example. A buyer needed about 300 cold-cast resin figures, 80 mm tall, with a 12 mm pin that had to fit a base. The shop poured a two-piece platinum RTV cavity against a CNC master, pips on the split, fiberglass jacket. The RTV itself measured near-zero after a weekend cure. The resin coupon shrank enough that the pin was recut on the master by a few tenths of a millimetre after the first ten casts. A parallel quote that applied a heat-cured LSR catalog percent to the master would have made the pin undersize in the opposite direction.

Put the method on the PO: production-compound coupon of the *cast* article, same dump as SOP, measured at 23 °C. Platinum RTV mold near-zero after full cure; tin RTV time-dependent at a stated age. Leave the master CAD-safe on critical logos until the first casting confirms direction. Freeze RTV lot family, mix ratio, and dump recipe — a new pectin or a new resin is a new coupon if the fit is real.

When this shrink advice does not apply

If you're buying vulcanized silicone parts (gaskets, keypads, overmolds) in heat-cured LSR or HCR, the cavity steel *does* take a heat-cure factor. Don't use this RTV stack on that RFQ — quote a parts OEM. A one-off art casting in tin-cure where nobody will recast also doesn't need a library coupon; freeze a wait and live with it. Geometry can override quantity: a hollow chocolate on a mandrel is two measurement states even at 80 dumps.

Related: what tooling means on an RTV PO, sampling the first dump, custom RTV molds.

Shrink note to put on the RFQ

Copy this. A factor without these lines isn't comparable.

• Platinum or tin RTV, mix ratio, degas, cure
• Dump/pour compound (pectin, chocolate, PU, wax) and working temperature
• Measure the *casting* at 23 °C after wait ____ (tin: also mold age)
• Master compensation for the cast article only — no LSR 2–3% class number
• Locked dims (core, jacket) vs free faces
• Who keeps the unsanded master for the next recast

Send the STEP, the dump compound, and whether anyone already applied an LSR percent to the master. We'll mark which layer actually needs compensation and whether the master should stay CAD-safe until the first coupon.

We would not recommend an LSR percent on a platinum master

We would not recommend pasting a heat-cured LSR 2–3% class factor onto a platinum RTV glove mold. That oversizes the master in the wrong direction. We would not stack CTE×ΔT on top of a coupon that already cooled, and we would not sand the master “to clean it up” if you expect a recast next season. Raising Shore A first is the wrong fix for a size miss; fix the dump coupon and the jacket.

RUUIPON Mold Structure Decision (shrink layer): platinum RTV after full cure is near-zero — compensate the *cast* article, not the gum. Tin is time-dependent — freeze a wait, or don’t treat it as a library. Hollow dump on a mandrel? two measurement states (locked ID vs free face). Soft Shore 10A–20A without a jacket? sag looks like “bad shrink.”

flowchart TD
  A[Which layer needs compensation?] --> B{Object on the RFQ}
  B -->|LSR/HCR molded part| C["Different quote: heat-cure 2-3% class on steel"]
  B -->|RTV cavity| D{Platinum or tin?}
  D -->|platinum, full cure| E["RTV ~0.1-0.3% — coupon the DUMP"]
  D -->|tin| F["Time-dependent: name mold age"]
  E --> G{Jacket / core lock?}
  G -->|free dump| H[Near-isotropic coupon at 23 C]
  G -->|mandrel or slumping glove| I["Two states: locked vs free; add mother mold"]

Typical shop values

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

| Item | Typical shop value | | --- | --- | | Platinum RTV linear shrink after full cure | About 0.1–0.3% | | LSR *part* contrast (wrong RFQ) | About 2–3% heat-cure class — do not put this on the master | | Mix | Platinum 1:1 or 10:1 by grade; vacuum degas until the foam collapses | | Cavity Shore | 10A–20A skin with a rigid jacket | | RTV wall / jacket | Skin often 8–15 mm; unsupported edge sag about 0.5–1.5 mm | | Coupon state | Casting at 23 °C after a named wait | | Outer-row gummy vs center (if tray sags) | Edge often 3–6% heavy vs the center pocket |

FAQ

Does platinum RTV share a shrinkage factor with LSR because both are platinum-cured? Platinum RTV mold rubber typically shows near-zero linear change after a complete room-temperature or mild oven cure. Heat-cured LSR for molded *parts* shrinks as it cools from cavity temperature and as post-cure drives off volatiles. Using an LSR catalog percent on a platinum glove mold oversizes the master. That LSR quote is a different product line.

Can I use the TDS shrink value if I match hardness and brand? Hardness isn't a shrink coupon. The TDS plaque used a lab cycle, a plaque thickness, and often a natural grade. Your wall, jacket constraint, dump compound, and wait time will move the number. Use the TDS only as a planning band until a production coupon exists.

How do inner cores and jackets change the factor? Constrained dumps can't shrink freely. A mandrel locks local dimensions while free faces still pull. Expect distortion, not a uniform scale. Design the core and the free section as two measurement states, and fixture the first article the way the part will live.

Is tin-cure RTV usable if I wait a week and then measure? You can freeze a *process* that way for a short run: pour, wait a defined time, measure, cast. You can't treat that measurement as a library dimension months later. For fits that must recast next season, use platinum RTV or accept a new master.

What temperature should shrink inspection use? Name it. 23 °C is the usual rubber-lab state. Measuring a dump still warm from a chocolate table, or a resin still exotherming, mixes CTE with shrink. Write “measure at 23 °C after wait X” on the drawing and on the first-article pack.