How to Cut Custom RTV Silicone Mold Development Cost

If you're trying to reduce custom RTV mold development cost, buy fewer recasts, not a starved pour. Frozen CAD, a cavity count that matches a real dump band, and sampling loops that change dump settings instead of the master will do more than a generic "make it cheaper" note on an RFQ. Skipping the jacket, skipping the witness dump, or substituting poured RTV for LSR to save a line item is how the second PO exceeds the first — and the LSR line was a different WYD-style *parts* quote anyway.

This is a cost-control sequence for a poured cavity. It doesn't publish a price. It doesn't tell you to pick the lowest China bid without matching process.

Freeze geometry before anyone mixes A/B

Every open dimension is a future master hour. Freeze parting, logo, texture, and wall of rubber before the box is built. Pay for a short DFM pass if the file isn't ready. That pass is cheaper than a recast, even when the DFM invoice feels like delay. A marked parting screenshot, a shrink owner, and a named process (platinum RTV or tin RTV) are the whole pass. If a bidder skips DFM and goes straight to a low number, they aren't cheaper. They're unpriced recasts.

If the article will change after a trade show, don't buy production cavity count. Buy a prototype path: a 1–2 cavity platinum RTV jacket for shape and peel. Write what that path is allowed to prove. Shape-only RTV shouldn't be asked to prove LIM flash.

China shops will pour an unfrozen sketch on request. That isn't a discount. That's an early mix date you'll pay for twice. Put freeze on the PO as a gate, with a revision letter.

Borrow process, not cavity, from a previous tool

Reusing a process window (same platinum family, same Shore skin, same dump) saves compound trials. Reusing a cavity that wasn't your shrink, not your split, and not your logo depth saves nothing after a plug and recast.

Family tools look like reuse. They work when hardness, shrink, and dump cycle can be shared. They fail when a shallow logo tile shares a sheet with a deep undercut mascot. That failure is a development cost even though you "saved a tool."

Insert reuse is the honest version: a standard jacket or mold box with a new RTV skin against a new master. Ask whether the base is yours. If it's the shop's, you reduced this PO and increased the cost of leaving.

One or two cavities as a decision, not a compromise

A one-cavity platinum RTV tool is often the cheapest way to get a real peel in the grade you'll sell as a tray. A two-cavity sheet can de-risk fill-weight before you cut sixteen pockets. These are decisions. They aren't shame.

Where they are the wrong save: a depositor nest you already know you'll run at 24 pockets on a frozen mascot; a jacket so floppy the one cavity already sags 0.5–1.5 mm. Jumping to 16-cavity while the lip is still in debate is the expensive personality.

Poured RTV is the right save when the silicone is the mold or when you only need shape. Using RTV to skip LIM development isn't a reduction. It's a postponed LSR *part* tool plus a pile of misleading samples. Using steel to "save later recasts" while you still dump chocolate is the same error in reverse.

Cut features that don't ship

Fine serifs, extra splits, decorative undercuts that the user never sees are development cost. Pad-print a logo if molded-in depth isn't the product. Simplify the mascot hair if a 10A–20A skin will tear at the root. Texture on the master also costs finishing and peel; skip grain the user will never see.

Paperwork bloat is similar. Don't buy a full automotive PPAP for a chocolate tray if you won't read it. Do buy the COC, first-article fill-weights, and the named food-contact file (FDA 21 CFR 177.2600 or LFGB) if you'll sell food-contact. Extra certificates you can't map to a grade aren't savings; they're clutter you still paid to photocopy. Don't invent SGS numbers to look cheaper.

Sampling loops that burn money without changing the part

Cap the number of CAD-driven sample loops in the quote. Process loops (dump temperature, jacket clamp, dwell) should be cheaper and should be used first. Cosmetic dislike without a datum is an infinite loop.

Ship samples with a written criterion: which fill-weights, which peel class, which color method. If the VP's opinion is the criterion, budget another loop. That's honest. Pretending the first tray will enchant everyone is how development cost hides in "just one more sample."

A practical China evaluation uses the same cap. Fast WeChat photos aren't free development. Each photo round that changes CAD is a recast with better lighting. Reduce cost by deciding, not by asking the shop to be slightly cheaper on an undefined object.

Also stop paying for process mix-ups. A quote that swaps poured RTV for LSR to hit a number will require the LSR tool later, plus the week you spent arguing about why the sample doesn't match production flash. A quote that swaps LSR steel onto a chocolate RFQ is the same waste. Name the process on the RFQ and refuse substitutes. That single line prevents the most expensive false save in this category.

Common mix-up: more cavities as a development discount

A lot of RFQs treat 16 pockets as "cheaper per dump, so cheaper development." In the shop that usually means you recast sixteen faces when the mascot moves 0.4 mm. Extra cavities reduce unit cost later if they fill evenly. They raise development cost now.

What to ask when you're buying

• Is this quote RTV mold-making, or HCR/LSR *parts*?
• Frozen CAD? If not, cap cavity count at one or two.
• Recast of the skin vs new master: which is priced?
• Cap on CAD-driven sample loops?
• Jacket in the price, or will sag show up as a "free" recast?

Worked example: 2-cavity first, then 16

A gummy line wanted a 24 mm fruit, 3.5 g target, food-contact, about 6,000 dumps a month after freeze. Development cost dropped by pouring a 2-cavity platinum skin (10A–20A, mix 1:1 or 10:1, rigid jacket) until fill-weight and peel froze, then recasting a 16-cavity sheet against the same unsanded master. Jumping to 16 on an open CAD, or to LSR steel "for scale," would have bought the mill or the pour twice.

When this cost advice does not apply

If you actually need thousands of molded silicone *parts*, the cheap RTV jacket isn't a discount; it's the wrong object. Quote HCR or LSR on a parts OEM. A tiny shallow one-piece ice tray can skip a heavy jacket. Geometry can override quantity: a three-piece core for a 45 mm hollow chocolate at 200 pours a month can cost more than a simple 12-cavity shallow tray at 4,000 dumps — and that's still the right spend.

Related reading: how much molds cost, prototype vs production, lead time, custom RTV molds.

Cost-control lines to copy onto the RFQ

Copy this. A lower number without these lines isn't a save.

• Process: platinum or tin RTV (ban HCR/LSR substitution)
• Freeze gate and revision letter
• First tool: 1–2 cavities until drawing freezes
• Jacket material in the price
• Cap on CAD-driven recasts; process loops cheaper
• Recast of the skin vs new master: two prices
• Witness pack: fill-weight, not a pretty photo

Send the pocket STEP, monthly dumps after freeze, and whether the mascot can still move. We'll price a 1–2 cavity peel check versus a production sheet so the cheap line is the same object.

We would not recommend starving the jacket to hit a number

We would not recommend RTV as a fake LSR production tool, or LSR steel as a fake chocolate tray. We would not buy 16 cavities on an open CAD. We would not raise Shore A first to "skip the mother mold." We would not skip the witness dump to save a report.

RUUIPON Mold Structure Decision: price the split and the jacket as hardware. Flat back? open-back. Deep undercuts that won't peel? two-part / plug / core — that's cost, not optional styling. Soft 10A–20A or wet dump? mother mold. Geometry overrides quantity.

flowchart TD
  A[Want lower development cost?] --> B{CAD frozen?}
  B -->|no| C[1-2 cavity RTV only]
  B -->|yes| D[Cavity count vs dumps per month]
  A --> E{Object is RTV cavity?}
  E -->|no| F[Stop: different part-tool RFQ]
  E -->|yes| G[Jacket in price; cap CAD recasts]

Typical shop values

Caption: Typical shop values (not a named lot).

| Item | Typical shop value | | --- | --- | | Mix | Platinum 1:1 or 10:1; vacuum degas | | Cavity Shore | 10A–20A skin with jacket | | Platinum shrink | About 0.1–0.3% after full cure | | Loaded-tray edge sag without a jacket | About 0.5–1.5 mm — a recast you didn't buy | | First tool | 1–2 cavities until freeze | | 24 mm fruit, then 16-cavity sheet | About 6,000 dumps/month after freeze is a common band |

FAQ

Is the cheapest way to reduce mold cost to use RTV instead of steel? Only if you needed a flexible cavity or a shape prototype. RTV doesn't replace LSR LIM or HCR compression. Using it as a fake production *part* tool defers the real tool and can mis-teach shrink and flash.

Will more cavities reduce development cost? No. More cavities reduce unit cost later if they fill evenly. They raise development cost now. Keep one or two cavities until the drawing freezes.

Does skipping first-article inspection save money? It saves a report and spends a recast or a failed lot. First article is how you stop paying for the wrong shrink or a sagging jacket.

Can I reuse another customer's tool to save development? Not unless you have the right to that cavity and it matches your process and shrink. Borrowing a catalog pocket is decoration. Borrowing a process window (grade, mix, jacket style) is the real save.

What RFQ line actually lowers China tooling quotes without hiding risk? A frozen CAD, a named RTV process, a 1–2 cavity first tool, a jacket in the price, and a cap on CAD-driven sample loops. Asking only for a lower number, with RTV/HCR/LSR left open, produces a bid you can't use.