FAI for Custom RTV Molds: What First Article Shows

If you're approving a custom RTV silicone mold, pull the packet apart before you look at the dump.

First article on a custom RTV tool exists to prove that *this* master, *this* platinum or tin skin, *this* jacket, and *this* trim made a cavity that matches the drawing and a dump that matches the fill-weight. It is not a capability study (that is later). It is not a PPAP unless you asked for one. It is not an AS9102 form unless you are in that system and you specified it. Most food, craft, and industrial RTV jobs need the substance of FAI, not an aerospace form number.

LSR part FAI against a model that was tuned on RTV is a fake pass. Don't mix them.

What the packet has to contain

Build it in this order. If a shop skips a layer, the rest is decoration.

1. The drawing you will hold them to, ballooned. Every dimension and note you care about gets a bubble number. Balloon flash max at the split, Shore of the cavity, fill-weight or pocket volume, post-wait of the dump, and the fits that actually mate. If you send a pretty CAD PDF with no balloons, you will get a “FAI” that measured five easy outsides.

2. Material identity. RTV grade and manufacturer, lot/batch, mix 1:1 or 10:1, hardness target, platinum vs tin. A cert that says “silicone, RoHS” is not a material cert. For FDA 21 CFR 177.2600 or LFGB claims, the cert must match the *cavity* compound, not the factory’s wish.

3. Process state of the measured dumps. As-cast or after a named wait, at what temperature (23 °C is the usual lab state). Platinum shrink is ~0.1–0.3%; tin is still moving. Measuring the wrong state is a fake pass.

4. Dimensional / fill-weight report against the balloons. Method named: scale, optical, caliper, water volume. Soft Shore 10A–20A wants contact strategy, not a macho CMM program copied from a steel housing. Record actuals, every pocket, not pass/fail ticks only. Outer-row 3–6% heavy vs center is sag.

5. Visual standard. Flash, voids, inhibition stick, jacket registration, keys. Photos with the same lighting you will use in incoming.

6. Tool identity. Master rev, jacket number, cavity count. An 8-cavity FAI that reports one pocket is a 1-cavity FAI with seven strangers.

That is the packet. Everything else (full capability, GR&R, biocompatibility reports) is extra and must be named at RFQ.

Common mix-up: photo-only FAI of the hero pocket

A passing photo of a nice bear does not prove pocket 8. Pull from a continuous short run, not the three prettiest dumps in a bowl.

ISO 3302-1 on the balloons is a drawing class if the article is rubber, not a substitute for a fill-weight map.

How to run it on the bench

1. Freeze drawing rev and balloons before the master is cut or printed. FAI against a moving rev is theatre.
2. Complete tryout until the dump window is boring: mix, vacuum, jacket, peel.
3. Pull first-article dumps from a *continuous* short run.
4. For multi-cavity, pull from every pocket. Report every pocket. If a pocket is blocked, the FAI says so.
5. Measure after the agreed wait and after the agreed trim. If you will ship untrimmed, measure untrimmed.
6. Hardness on a thick enough section of the *skin*, or on a companion slug, because a 1 mm lip will lie on a durometer.
7. Pack the packet with the mold, cavity-tagged, plus master custody note.

If fill-weight fails, stop. Don't “FAI the next lot.” Either the jacket, the master, or the pour is wrong. Recast or re-support, then new first article. Partial waivers belong on a signed deviation, not in email.

Silicone-specific traps

Softness. A Shore 10A skin will move in a vise. Fixturing and low-force probes are part of the method.

Flash as a dimension. If flash max is 0.15 mm at the split, it is a balloon, not a vibe.

Shrink vs LSR samples. Platinum RTV dumps are a different shrink story than heat-cured LSR. Do not FAI an RTV tray against a model that was tuned on LSR steel, or the reverse.

Sag. Straightedge on the jacket back. 0.5–1.5 mm edge dish is a missing mother mold.

Inhibition leftovers. Sticky first articles on platinum are a chemistry fail. They don't belong in the dimensional table.

Family tools. Each unique pocket is its own first article, even if the PO thought it was one dump.

What to ask when you're buying

• Balloons with the RFQ, not after tryout?
• Sample size: often 1–5 per pocket?
• Ship state: wait, trim, temperature?
• Every cavity reported?
• Form: shop table vs your template vs a named aerospace standard *if you actually need it*?

Worked example: 12-cavity bear packet

A 12-cavity 18 mm bear, 4.0 g target, food-contact, ~2,000 dumps a month. Balloons: fill-weight, flash at split, Shore of skin, chin radius. Platinum 10A–20A, mix 1:1, vacuum degas, jacket. FA: twelve weighs, straightedge on jacket, photos of keys. Hero-pocket photo was refused. A parallel LSR FAI form with gate remnant balloons was the wrong packet.

When this article is the wrong quote

If you actually need FAI of vulcanized LSR/HCR parts, gate remnant, post-cure, and cavity ID in steel belong on a parts OEM pack. A one-off artisan dump can skip aerospace forms. Geometry can override quantity: a family tray is still N unique FAI at 80 dumps.

Related reading: custom RTV molds, DFM checklist, turnaround times, gummy trays.

What FAI does not prove

A passing FAI does not prove 10,000 dumps. It proves the recipe *can* hit the drawing. Capability and incoming sampling are separate. A printed master can pass first article on pour 12 and dent by pour 200. If life matters, put a short run and a jacket photo in the packet, not just one piece.

Buyer actions: send balloons with the RFQ; state the form; book incoming to the same balloons. If your IQC uses a different drawing, you bought two truths.

End the PO note with: “First article = ballooned drawing + RTV cert (grade/lot) + fill-weight/actuals per balloon per pocket + visual photos + jacket/master ID, measured in the ship state (wait/trim yes or no). No photo-only FAI.”

Send balloons and dump material. We will run every pocket and tell you if the jacket, not the CAD, failed.

We would not recommend photo-only FAI of the hero pocket

We would not recommend measuring a dump still warm from the table. We would not FAI LSR against an RTV-tuned model. Raising Shore A first is the wrong fix for a fill-weight miss; fix the jacket. We would not skip a blocked pocket in the report.

RUUIPON Mold Structure Decision: FAI the structure you bought. Flat back? open-back still gets a straightedge. Deep undercuts? peel photos. Soft 10A–20A or wet dump? mother mold in the packet.

flowchart TD
  A[FAI packet?] --> B{Balloons frozen?}
  B -->|no| C[Stop: theatre]
  B -->|yes| D[Every pocket: weight + visual]
  D --> E{Jacket flat?}
  E -->|no| F[Fail: sag, not CAD]
  E -->|yes| G[Cert + Shore + ship state]
  A --> H{LSR part FAI?}
  H -->|yes| I[Stop: different packet]

Typical shop values

Caption these as 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%; tin still moving — name wait | | Measure state | 23 °C after a named wait | | Loaded-tray edge sag | About 0.5–1.5 mm without a jacket | | Sample size | Often 1–5 dumps per pocket |

FAQ

Is a photo of a nice dump an FAI? No. FAI is a packet: ballooned drawing, RTV identity with grade and lot, process state, fill-weight/actuals against those balloons, visual photos, and master/jacket ID. It is not a capability study, not PPAP unless you asked, and not AS9102 unless you specified that system.

Why must the measured dumps name wait versus as-cast? Platinum is near-zero after full cure; tin moves; dumps still warm mix CTE with shrink. Measuring the wrong state is a fake pass. Measure after the agreed wait and after the agreed trim.

Does a green FAI prove the next 10,000 dumps? It proves this master, skin, jacket, and trim can hit the drawing. Capability is separate. A printed master can pass on an early pour and fail later; if life matters, put a short run and a jacket photo in the packet, not one pretty piece.

Why is an 8-cavity report on a single pocket incomplete? Each pocket is its own first article. Pull from a continuous short run, not the three prettiest dumps in a bowl, and report every cavity. A blocked pocket must be stated. Family tools are unique pockets even if the PO thought they were one dump.

What PO note kills photo-only FAI? First article equals ballooned drawing plus RTV cert (grade/lot) plus actuals per balloon per pocket plus visual photos plus jacket/master ID, measured in the ship state. Send balloons with the RFQ, not after tryout, and book incoming to the same balloons.