Calculating ROI on Upgrading from Single-Cavity RTV to Multi-Cavity LSR Molds

Single-cavity RTV vacuum casting and multi-cavity LSR injection are not two sizes of the same mold. One is a technician with a platinum or tin jacket. The other is a hot-cavity, cold-runner press with cavitation. ROI is whether the second factory pays for itself *before the product dies*, on drivers you can name.

Do not compare “mold price” to “piece price.” Compare a horizon (12 or 24 months is enough for most custom parts) of labor, scrap, material form, setup, and the recut you will do if CAD is not frozen. Invented dollar paybacks are how teams buy eight cavities for a geometry that will change in March.

Name the two factories

Leaving: an RTV tool (usually platinum-cure jacket around a master; tin-cure is cheaper and less stable). Cycle is mix, degas, pour or inject by hand, cure, demold, trim. Shrink and flash behave like casting. Chemistry is often a stand-in for the production LSR. Jacket life is counted in casts, not in SPI class.

Buying: LSR on metal, typically an aluminum 1-cavity bridge or a steel 2-/4-/8-cavity production tool with cavity heat and a cold runner. Cycle is clamp, inject, cure, eject. Flash is a shutoff. Shrink is production shrink. Grade should be the production platinum LSR.

There is a missing middle: aluminum 1-cavity LSR. Many “RTV to 8-cavity steel” ROI models skip it and then discover LSR fill, vestige, and shrink on the most expensive tool. Put the middle in the model even if you skip it in the buy.

Drivers, not stickers

Write one line each. Ranges of *structure* are allowed. Fake unit prices are not.

T_tool. RTV: master + jackets, recurring as jackets die. LSR: class intent (105/104/103…), cavity count, cold deck vs open sprue, vacuum, slides, tryout, FAI. Multi-cavity is a step-change in plates, balancing, and tryout, not “N times a single cavity.”

T_mod. First geometry change. RTV makes this small. Hardened multi-cavity steel makes this large. If CAD is live, T_mod is not zero.

C_labor. RTV cell time per good part (mix through pack). LSR: tend, destage, inspect. Cavitation cuts press occupancy; it does not automatically cut destage if every cavity flashes.

C_mat. Mass × grade × (1 + runner/jacket waste). RTV waste is cups, sprues, and failed pours. LSR waste is start-up, cured sprue if you refused a cold deck, and purge at color change. Production LSR gum is often a different cost structure than the RTV casting compound. Put both grades on the sheet.

C_scrap. RTV scatter, bubbles, jacket wear. LSR start-up shots, imbalance on new multi-cavity tools, post-cure mishaps.

C_setup. Per batch. Multi-cavity raises MOQ thinking. A beautiful annual volume that arrives as 200-piece releases puts setup back on RTV’s side.

C_quality. The cost of a part that *looks* like production but is not: wrong shrink, wrong compression set, wrong extractables. This is the driver that kills “RTV is cheaper” when the product needs LSR chemistry now.

Horizon H: T_tool + T_mod + Σ(C_labor + C_mat + C_scrap + C_setup + C_quality events).

ROI is the *delta* on those C terms versus the *delta* on T_tool and T_mod. If you cannot fill C_labor from a timed RTV cell, you do not have ROI. You have a slide.

A crossover you can run without fake prices

1. Time the current RTV cell on a normal lot. Convert to hours per 1,000 good parts.
2. Get two LSR constructions quoted as *stacks*, not as piece prices: (A) 1-cavity aluminum LSR, (B) steel at the cavity count you think you want. Same grade, same flash spec, stated cycle, stated runner waste, stated scrap assumption, stated batch size.
3. Freeze H and freeze batch size. Annual quantity without batch size is a lie.
4. Ask whether chemistry is already LSR. If the product requires a healthcare, self-lubricating, or specified platinum LSR, RTV stand-in parts do not belong in the “savings” column. Switch process first, cavity count second.
5. Apply T_mod: number of drawing changes expected in H. Each change on steel is a recut class. Each change on RTV is a new jacket or a patched master.

The upgrade wins when labor hours plus scrap plus chemistry-mismatch risk in H exceed the LSR tool stack, *and* the CAD is frozen enough that you will not pay for the steel twice.

The upgrade loses when volume is real on a spreadsheet, geometry is still competing, and you skipped aluminum.

When multi-cavity is the expensive part of the sentence

Cavitation pays when cure time dominates and the press is the constraint: more cavities, same heat, more parts per hour, *if* yield holds. It does not pay when:

• Annual volume will not keep that press fed.
• Releases are small; you will still pay setup and start-up scrap on a large tool.
• The part is unbalanced (thick/thin) and 8 cavities will teach you 8 ways to flash.
• You have not run even 1-cavity LSR, so shrink and gate vestige are unknown.

A common failure is treating “single-cavity RTV → multi-cavity LSR” as one decision. It is two: process change (casting compound → LSR on metal) and capacity change (1 cavity → N). Process change can have positive ROI at hundreds of parts if labor or chemistry demands it. Capacity change needs a frozen drawing and a real annual quantity.

Data the quotes must include

• Construction: RTV jacket count vs LSR class, cavities, runner type.
• Cycle or cell time, and whether cure dominates.
• Runner or pour waste as a fraction of shot or pour mass.
• Trim method.
• Scrap assumption for start-up and running.
• Setup per batch; quoted MOQ.
• Stated PM / jacket replacement interval.
• Grade: name the LSR and the RTV compound separately.

If they only give a unit price at 10,000, you cannot calculate ROI. You can only pick the number that looks smaller.

Decision test: fill the driver sheet for stay-on-RTV, aluminum 1-cavity LSR, and steel N-cavity. The lowest sum over H wins. If aluminum is the lowest and you still jump to eight cavities, you are buying a monument for a meeting.

FAQ

At what annual volume should I leave RTV for LSR steel?

There is no universal quantity. Leave when timed labor, scatter, or production chemistry already beat a 1-cavity LSR stack on your horizon. Tens of thousands of a frozen part is usually steel; hundreds may still be RTV if the stand-in compound is acceptable.

Is aluminum 1-cavity LSR a better ROI step than jumping to 8-cavity steel?

Often: Aluminum teaches fill, vestige, shrink, and flash on a tool you can recut. Eight-cavity steel teaches those on the tool you cannot afford to recut twice. Put both constructions on the same driver sheet. Skip aluminum only when grade, geometry, and annual quantity are already boring.

Does a cold-runner multi-cavity always win on material?

No: A valve-gated cold deck cuts cured sprue waste; it also raises T_tool. On a large industrial part with a hidden gate and modest volume, an open cold sprue can win. Run waste as a fraction of shot weight against deck complexity. Tiny parts where the tree outweighs the rings are the classic deck win.

How do I treat jacket replacement in the RTV column?

As recurring T_tool, not as zero. Platinum jackets die from tear, stretch, and inhibition; tin jackets die faster and with more dimensional drift. Count jackets per horizon. Ignoring them makes RTV look cheaper than the cell you actually run.

What batch size should I freeze in the model?

The release quantity you will actually pull, not the annual forecast. Multi-cavity economics assume the tool stays in the press, so 200-piece drops can erase cavitation with start-up scrap. If only the pretty release wins, you do not have ROI; you have a forecast.