How Many Cavities Should Your Injection Mold Have?

Cavity count is the first number in a mold specification and usually the least examined. It gets set early, often by whoever remembers what the last similar tool had, and it then determines the tool cost, the press, the mold weight, the runner, the tolerance the tool can hold, and how many parts you scrap when something goes wrong. Very little of that is reconsidered afterwards, because by then the steel is cut. This page is about getting the number right on paper.

The short answer, and the four numbers behind it

For most production parts the answer falls between two and eight. Below twenty thousand parts a year, one cavity is usually correct. Above a million, it is rarely fewer than sixteen. In between sits the whole argument, and it is settled by four numbers rather than by preference: the annual volume, the peak month, the largest press you can actually put the tool in, and how much of the tool cost the volume can repay.

One term worth clearing up first, because it causes confusion in quotations. Some shops use cavitation to mean the number of cavities; others use it to mean the voids that form in a thick section when packing is inadequate. This page is about the first meaning. If a supplier’s document mentions cavitation and you are not sure which, ask — the two have nothing to do with each other and both appear in the same reports.

There is one more thing to settle before the numbers. Cavity count is not a production decision, it is a capacity decision, and capacity has a shape that unit economics does not show. A tool with four cavities cannot make one part; it makes four or none. That single fact is why high-cavity tools need a different kind of planning, a different kind of maintenance, and a different answer to the question “what happens if a cavity is damaged”.

The arithmetic that actually decides it

The instinct is that doubling cavities halves the cost per part. It does not, and understanding why is the whole argument. The cost per part has three components, and only one of them is divided by the cavity count.

The machine term is divided by cavities, but the machine grows. The press is charged by the hour, and its hourly rate is fixed by its size. Doubling cavities doubles the output from the same hour — but it also doubles the projected area, which doubles the clamp force required, which means a bigger press at a higher hourly rate. In practice the press rate rises more slowly than the tonnage, because a 600 t machine is not twice the cost of a 300 t one, so there is a real saving. It is just smaller than half.

The tool term is multiplied by cavities, and it grows roughly in proportion. A four-cavity tool costs considerably more than a single-cavity one. That extra cost has to be repaid out of the parts, and it is repaid entirely from the saving in the machine term.

The material term is barely affected at all. Every part needs the same shot weight of plastic; adding cavities does not reduce it. What it does change is the runner: in a cold runner tool the runner weight grows with cavity count, and at high cavity counts the runner can become a substantial fraction of the shot. That scrap is either reground or thrown away, and both cost something. How the shot behaves in the barrel and in the cavity is the subject of the injection molding process; for cavity count, the point is only that material is the one cost that does not divide.

The table below isolates the first term, using a 150 × 100 mm part with a 2 mm wall, ABS, a 25 s cycle. Press rates are indexed to a standard 80 t machine at 100 and scaled by the square root of tonnage, which is close to how machine rates actually behave. Cavity and runner areas are approximate and rounded.

Two things in that table are worth staring at. The first is that the press cost per part falls steeply at the start and then flattens: halving it takes going from one cavity to four, and then it takes another twelve cavities to halve it again. The second is the last column of the last row: thirty-two cavities needs a two and a half thousand tonne press. Almost no molder has one, and the ones that do are not running a 150 mm part on it.

So the honest conclusion is not “more cavities are cheaper”. It is that the saving from more cavities is real, front-loaded, and blocked by something other than arithmetic — by the press that exists, by the tool cost that has to be repaid, and by the physical size of the thing you are trying to lift. Those three are the real constraints, and they are what the rest of this page is about.

Cavities multiply the press, and the press is the hidden cost

The clamp force calculation itself is straightforward — projected area times cavity pressure, with a safety factor — and it is set out properly in clamp force and tonnage. What matters here is the consequence for cavity count, which is that projected area is linear in cavities. Every cavity you add adds its whole footprint to the force the press has to hold shut, including the runner that feeds it.

This produces a pattern that catches people out. A tool that is perfectly reasonable at four cavities becomes a different project at sixteen: the press roughly doubles in size, the mold base grows two frame sizes, the tool weight roughly triples, and the crane at the molder may no longer be able to lift it. None of those are visible in a cost-per-part calculation, and all of them decide whether the tool can be run at all.

The shot capacity constraint is the one most often missed. The injection unit has to deliver the whole shot — all cavities plus the runner — and best practice is to keep the shot between roughly twenty and eighty percent of the barrel capacity. Below twenty percent the melt residence time becomes long enough to degrade heat-sensitive resins; above eighty percent there is no cushion left for consistent packing. This puts a ceiling on how many cavities can be added to a given machine’s injection unit independently of the clamp force, and it is not unusual for the injection unit to be the binding constraint rather than the clamp.

And then there is handling. A sixteen-cavity tool for a 150 mm part is a heavy object. It has to be lifted by hooks in designated holes, moved on a crane rated for it, loaded into the press with a hoist, and stored on a stand that will not distort the base. Shops that buy a high-cavity tool without checking the crane capacity discover the problem on delivery day, which is the worst possible day to discover it.

Matching cavities to volume

The practical way to choose is to work backwards from the peak month and the available press time, not forwards from the annual total. The table below is a starting point, not a rule — a large part shifts the answer down and a small part shifts it up, because press size and tool cost move in opposite directions with part size.

Two adjustments matter more than the table itself. The first is the peak month. A product with a seasonal peak that is three times the average needs capacity for the peak, and capacity that sits idle for nine months is expensive. The standard answer is to size for something between the average and the peak, and cover the difference with a second shift or a second tool. The second is the number of shifts you are willing to run. A single-cavity tool on three shifts produces three times what it produces on one shift, and if the molder is willing to run nights, that often beats buying cavities — because it costs nothing up front.

The interaction with cycle time is worth stating plainly, because it is where most cavity-count arguments go wrong. Output is cavities divided by cycle time. Adding cavities and shortening the cycle are two ways to buy the same output, and they have very different cost structures: cavities cost tool money, cycle time costs process engineering. When someone proposes doubling the cavities to solve a delivery problem, the first question should be whether the cycle is anywhere near where it should be — because a tool that is running at 35 seconds when it should run at 22 has forty percent of free capacity already, and cooling is usually where it is hiding.

Runner balance is where multi-cavity tools fail

A multi-cavity tool is a promise that every cavity fills identically, at the same instant, at the same pressure, and at the same temperature. That promise is kept by the runner geometry or it is not kept at all. Once the cavity count goes above four, the runner becomes the hardest part of the tool design, not the part geometry — which is why gate and runner layout sits near the top of any serious DFM review.

The practical consequence is a diagnostic that anyone can apply. If one cavity consistently short-shorts, or one cavity always shows a heavier weld line, the tool is not balanced — and the fix is in the runner, not in the process. Raising injection pressure until the weak cavity fills will over-pack the strong ones, which produces flash on some cavities and sink on others, and then the tool has two problems instead of one.

This is also where the concept of effective cavitation comes from. A sixteen-cavity tool that is badly balanced may only be running twelve cavities properly, with four filling late and producing parts that are dimensionally different and cosmetically worse. The tool was bought as sixteen and it is producing twelve, and nobody notices because the short shots are sorted out and scrapped rather than investigated. Checking this is straightforward: weigh the parts from each cavity, or measure one critical dimension on ten parts from each. If the spread across cavities is bigger than the spread within a cavity, the runner is the problem.

The choice between a cold runner and a hot runner interacts with all of this. A hot runner removes the runner completely, which removes the balance problem and the scrap at the same time — but it adds a manifold whose cost and lead time grow with drop count, and it introduces a new failure mode in every drop. The trade-offs are set out in hot runner versus cold runner; for cavity count purposes, the relevant fact is that hot runners make high cavity counts practical at the cost of a manifold that must be specified early.

Every cavity multiplies the risk, not only the output

This is the part of the argument that does not appear in a cost model and should. A multi-cavity tool does not produce sixteen identical parts. It produces sixteen parts with a measurable spread between cavity positions, and that spread is added to the repeatability of each individual cavity.

The row that matters most is the last one. The weakest cavity sets the process window for the whole tool. If one cavity has a slightly thinner wall, or a slightly longer flow path, or a marginally different gate, then the pressure needed to fill that cavity is the pressure the whole tool receives. The other fifteen are over-packed as a consequence, which shows up as flash on the parting line, higher internal stress, and slightly different dimensions than the weak cavity produces. This is why a sixteen-cavity tool often cannot hold a tolerance that a two-cavity tool holds comfortably, even when both are made to the same standard.

The mitigation is design, not process. Number the cavities and engrave the numbers where they can be read after molding, so a defect can be traced. Insist on a per-cavity capability study at the trial rather than a study on a mixed sample, and make the acceptance criterion explicit: not “samples approved” but “all sixteen cavity positions within tolerance, with the between-cavity spread stated”. What that trial should establish, and how to sign it off, is covered in mold trial approval, and the tolerance side in injection molding tolerances.

The four physical limits on cavity count

Above a certain point the argument stops being economic and becomes physical. Four limits do the work.

Clamp force. Projected area times cavity pressure, linear in cavities. This sets the press, and the press sets the floor on part cost.

Injection unit capacity. The shot, including the runner, must sit between roughly twenty and eighty percent of barrel capacity. This is often the binding constraint before clamp force is.

Mold size against the press envelope. The base has to fit between the tie bars and inside the platen, with enough daylight to open and enough stroke to eject. A high-cavity tool grows in both directions, and a press that has the tonnage may not have the platen.

Handling and storage. Weight against crane capacity, and floor space for the stand. This sounds administrative until a tool arrives that cannot be lifted.

Below all four sits a fifth, softer limit that is worth naming because it is the one that actually stops most projects: the number of cavities you can afford to lose at once. At one cavity, a damaged cavity is a stopped tool. At thirty-two, a damaged cavity is a stopped tool unless the design allows it to be blocked, and a blocked cavity in a balanced runner is itself a problem, because the balance was calculated for all thirty-two. This is why high-cavity tools are usually specified with the ability to run at reduced cavitation, and why that capability has to be designed in rather than improvised.

There is a sixth limit that belongs with the physical ones even though it is metallurgical rather than mechanical: consistency of the cavity steel. Sixteen cavities cut from one block behave alike. Sixteen cut from different lots can differ enough in hardness and machinability to show up as a dimensional spread that no process setting will remove. Specifying one grade, one hardness range and ideally one heat is part of mold steel selection, and it matters more as the cavity count rises — because a single reworked cavity has to match fifteen others that are already finished.

The alternative to more cavities: a second parting plane

There is a way to double output without doubling the press, and it is worth knowing about before the cavity count is fixed. A stack mold has two parting planes and two sets of cavities, one above the other, fed by a runner through the center. The clamp force is set by the projected area of one level, not both, because the two levels balance each other against the moving center plate. So a two-level stack mold produces twice the parts on the same press.

It is not free. The tool is taller, so it needs a press with more daylight and a larger stroke. The stack is more complex to build, roughly half again the cost of the equivalent single-level tool, and two to four weeks longer. Ejection, core pull and any sliders have to work on both levels, which is where most of the difficulty lives. And the tool weighs considerably more, which brings the crane limitation back.

For a high-volume part where the press is the bottleneck rather than the budget, a stack mold is often the correct answer and is routinely overlooked because it does not appear in the usual cavity-count discussion. For everything else it is unnecessary complexity. The judgement is simple: if a bigger press is available and affordable, more cavities is easier. If it is not, a second level is the way to get output out of the press you have.

Staging the decision

Cavity count is usually decided once, at the quotation, on the basis of a volume forecast that has not yet been tested. That is an uncomfortable place to make an irreversible decision, and it is why staged builds exist.

The second row of that table is the one worth understanding properly. Building four cavities into a base that has pockets for eight costs perhaps a third more than a plain four-cavity tool, and adds a week or so to the schedule — the timing of that is covered in injection mold lead time. What it buys is the ability to add the second set of four when the volume arrives, without buying a new base, a new ejection system, a new cooling circuit or a new hot runner manifold. If the volume doubles within about eighteen months, the staged route is clearly cheaper. If it does not, you have paid a third more for nothing.

The condition that makes staging work is that the part must not change. A base with pockets for eight cavities is only useful if the four cavities you add later have the same geometry as the four you built. If there is a reasonable chance the design will be revised, staging buys nothing, and the money is better spent on a smaller tool that can be replaced outright. For a part that is still moving, a single-cavity or two-cavity tool is not a compromise; it is the correct tool for an unsettled design.

Family tools, and when they are a mistake

A family tool molds several different parts in one shot, in one tool, from one material. The attraction is obvious: one tool instead of four, one setup instead of four, one press occupying time instead of four. For a set of small parts that always ship together, it can be the best decision in the project.

The condition is arithmetic, not preference. A family tool balances only when the parts share three things: the same material, a similar wall thickness, and a similar volume share of the shot. Material and wall determine the pressure and temperature the shot needs; volume share determines which cavity fills first. Get the volume share wrong and the tool behaves exactly as an unbalanced runner does — the small part flashes while the large part short-shorts, and there is no pressure setting that fixes both at once.

The failure mode to watch for is the slowly diverging demand. Two parts start with similar volumes and the tool is balanced for that. Eighteen months later one part is selling three times the other, and now the tool is producing the wrong mix — too many of the slow-moving part, not enough of the fast one, and both at the pace of the slow one because they share a shot. A family tool locks the production ratio of its parts together for the life of the tool. That is fine when the ratio is stable and expensive when it is not.

There is also a maintenance asymmetry worth knowing. In a family tool, a problem with any one cavity stops all the parts. In separate tools, it stops one.

Ten questions that settle the number

Before the cavity count is written into a purchase order, these ten questions are worth answering in writing. Each one can move the answer, and most of them are cheaper to answer now than to correct later.

Two of them deserve emphasis because they are routinely skipped. The peak month, because capacity is bought for the peak and paid for all year. And whether a cavity can be blocked off, because it decides what happens on the worst day of the tool’s life. A tool that can run at fourteen of sixteen cavities keeps the line moving while a repair is scheduled. A tool that cannot, stops the line for the duration of the repair.

Five mistakes that show up in real specifications

The last one is the most expensive and the most common. Cavity count and cycle time are two ways of buying output, and they are not interchangeable in cost. Adding cavities costs tool money, press money and a longer build. Shortening the cycle costs process engineering, and the free capacity is already there if the tool was designed properly. When a program is short of capacity, the honest first question is what the cycle time should be — and if nobody knows the answer, that is the more valuable investigation.

None of this makes cavity count a difficult decision. It becomes difficult only when it is made once, early, by habit, and then treated as fixed. The decision is really four numbers — volume, peak, press, payback — and a conversation about tolerance and cosmetic risk. Ten minutes of that conversation before the base is ordered is worth more than any amount of process tuning afterwards. If you want a second opinion on a specific part, send the drawing, the annual volume and the peak month along with any press constraints, and the cavity count, the tonnage and the tool shape can be worked out from those.

Frequently asked questions