Common Failures When Making Custom RTV Silicone Molds

If a custom RTV silicone mold is failing, name the chemistry first, then treat the defect as geometry or as poison — not as a mystery. Sticky platinum faces are contamination on the master or the pour. Incomplete copy and bubbles are degas, pour head, or a dirty SLA. Tears at splits and logos are draft, dry keys, or a missing jacket. Size fights are tin-cure age, a slumping glove, or someone who pasted an LSR shrink percent onto the master.

Unmix the job before you diagnose. An RTV pour that copied a dirty print isn't an LSR short shot. HCR underfill and cold-runner valves belong on a molded-*part* RFQ, not this cavity. If the PDF in your inbox is full of heaters and manifolds, you're on the wrong defect list.

Symptoms → confirm → fix

Sticky cavity, undercured skin. Platinum addition-cure poisons against sulfur, some tin-cure residues, certain latex gloves, amines, and a long list of UV-resin leftovers. Confirm: the face stays tacky after extra oven time. Fix the master (wash, UV post-cure, barrier coat), the handling, or the box cleanliness, then recast. Extra bake will not un-poison a surface.

Bubbles at the parting, bald texture. Incomplete vacuum degas, pour head too low, or grain deeper than the dump can fill. Confirm on mix-sheet vacuum time and a photo of the last-to-fill pocket. Repour that skin; raise head or add a vent in the mold box. Don't "inspect in" a bubble cluster.

Flash at the split / keys that don't shut. Dirt on the split, worn pips, a jacket that rocks, or a one-piece glove asked to do a two-sided logo. Confirm with a dry-fit rock test and key photos. Recast with keys, or add a mother mold. Adding clamp on a warped jacket is the usual wrong first move.

Tear at a letter or core. Sharp CAD, no stretch path, aggressive demold, inner core without a pull path. Confirm the nick location vs the draft. Recast while the master still registers; add draft or a split. If the core can't extract, you designed a cut mold.

Size drift next season. Tin-cure still moving, UV storage, stacked jackets, or a master that was sanded "to clean it up." Confirm retain dump vs new dump at 23 °C, and mold age if tin. Switch to platinum, store unstressed, keep the unsanded master.

Cavity 1 fills, cavity 8 is short. Pour-head scatter or jacket registration, not "bad rubber." Demand all-cavity first dumps. Balance the box; don't invent a second shrink factor.

Common mix-up: more clamp, more bake, more time

A lot of RFQs treat flash as a clamp problem and sticky platinum as an undercure problem. In the shop that usually means you're crushing a jacket that already rocks, or you're baking a poisoned skin. Inspect keys and the master first. Clamp isn't a mill. Oven time isn't an antidote.

Overseas lag is a protocol issue: cavity-tagged photos, mix-sheet revision, one file vault. China cluster shops can be excellent here when you send evidence. They can't read a voice note.

What to ask when you're buying (or diagnosing)

• Platinum or tin, and do we have the mix sheet (ratio, degas, lot, pour date)?
• Sticky, torn, short, or size — which symptom, which cavity, keys in the photo?
• Master still unsanded and in whose custody?
• Jacket rock test: does it locate without tape?
• Dump compound and temperature vs what the skin was poured for?

If the shop only says "we'll add clamp," you haven't started diagnosis.

When this defect list is the wrong list

LSR knit lines, heater zones, and cold-runner valves are a parts-OEM problem. Extrusion die swell doesn't belong on a cavity Pareto. A ten-dump tin art glove doesn't need a multi-cavity balance study. Geometry can override quantity: one locked mandrel ID will distort even at 50 chocolates.

Related: extend mold life, inspection on arrival, shrinkage stack.

Worked example

A 12-cavity gummy, 18 mm, food-contact platinum, keyed split, resin jacket, a couple thousand dumps a month. Cavity 6 went sticky and cavity 11 tore at the ear. Mix sheet showed a normal 1:1 pour; the master for 6 still had SLA inhibition (tacky after a weekend bake — chemistry, not undercure). They barrier-coated the master and recast 6. Cavity 11 was a sharp ear with no stretch path; they added draft on the master and recast. Extra oven on 6 would have shipped a poisoned skin.

Defect RFQ / photo pack to send

• Symptom + cavity ID + keys in frame
• Platinum/tin, mix sheet, mold age if tin
• Master photos (layer lines, residue) and jacket dry-fit
• Dump compound and temperature
• What you already tried (clamp, bake, cleaner)

Send those photos and the mix traveler. We'll say recast, reclean the master, recut draft, or new jacket — not a slogan about quality.

We would not recommend more clamp, more bake, more time as first moves

We would not recommend crushing a rocking jacket to kill flash, or baking a poisoned platinum skin to un-stick it. We would not invent a second shrink factor because cavity 8 is short. We would not send a torn tin mold back as a polish complaint. Inspect keys and the master first.

RUUIPON Mold Structure Decision: sticky = chemistry on the master, recast. Flash = keys or jacket, not clamp. Tear at a letter = draft or missing split. Sag 0.5–1.5 mm = mother mold. Hollow mandrel that will not extract = cut mold, not a handling trick.

flowchart TD
  A[Defect] --> B{Sticky after extra bake?}
  B -->|yes| C["Inhibition: reclean master, recast"]
  A --> D{Flash / keys open?}
  D -->|jacket rocks| E[New jacket / recast with keys]
  A --> F{Tear at letter or core?}
  F -->|no stretch path| G[Add draft or split, recast]
  A --> H{Cavity 1 full, 8 short?}
  H --> I["Balance box / vents — not a second shrink %"]

Typical shop values

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

| Item | Typical shop value | | --- | --- | | Mix | Platinum 1:1 or 10:1 by grade; vacuum degas until the foam collapses | | Cavity Shore | 10A–20A skin with a rigid jacket | | Platinum shrink after full cure | About 0.1–0.3%; tin age is a size-drift cause | | Loaded-tray edge sag / jacket rock | About 0.5–1.5 mm | | Outer-row gummy vs center if the tray sags | Edge often 3–6% heavy — looks like “bad rubber” | | RTV skin wall | Often 8–15 mm; thin letters tear first | | Production food tray | 12–24 cavities; demand all-cavity first dumps |

FAQ

Why does more clamp often make split flash worse over a week? You're crushing a jacket or keys that were already tired or dirty. First dumps may look tighter; then the split rolls and flash grows. Inspect and recast or re-jacket. Clamp isn't a mill.

Can I post-cure longer to fix a sticky platinum mold? Not if the face is inhibited. Inhibition is chemistry at the surface. Extra bake can harden a marginally undercured bulk pour; it won't convert a poisoned skin into a clean skin. Find the poison (master, insert, glove, residue) and recast.

Is cavity imbalance always a pour-head problem? Often, not always. A plugged vent on one pocket, a nicked key, or a rocking jacket can mimic imbalance. Check registration and vents before you recut a box that wasn't the cause.

Why do first dumps pass and season-two dumps fail? Common causes: tin age, new dump lot, UV storage, mixed cavities, measurement while warm. Treat the restart as a new sample until those match the retain. Retains from first dump are how you prove the drift.

Are RTV tears a design problem or a handling problem? Both. Undercuts without a stretch path and sharp corners are design. Dry, unstressed storage and avoiding solvents that swell the rubber are handling. Tin-cure aging is a material choice. Don't send a torn tin mold back as a polish complaint.