Cooling is the largest single block of an injection molding cycle — normally 60 to 70 percent of it, and more on a thin-wall part. It is also the part of the mold design that is reviewed last, specified out of habit, and trimmed first when the quotation looks high. That combination explains most of the tools that arrive with a cycle nobody can account for, and why the fix is almost never in the process settings.
Cooling is the cycle, and the cycle is the cost
A molding cycle has four parts: fill, pack, cool, and the mechanical time to open, eject and close. Fill and pack together are usually a few seconds. The mechanical time is a few more. Everything left over is cooling, and on a normal tool it is more than half of the total.
That asymmetry is the whole reason cooling deserves design attention rather than a habit. Every other part of the cycle can be shortened by pushing the machine harder. Cooling cannot, because it is limited by how fast heat can leave the plastic through the steel and into the water — and that rate is set by geometry that is already cut into the mold.
Which is also why cooling is a cost question wearing an engineering coat. Cycle time is the largest variable in what a molded part costs, so the water circuit in the tool is a direct input to what the mold and the part will cost — not a technical afterthought to be sorted out at the trial.
The governing relationship is simple and it is brutal: cooling time scales with the square of wall thickness. Double the wall and the cooling time does not double, it quadruples.
| Wall | Cooling time, relative | What happens to the rest of the cycle |
|---|---|---|
| 1.0 mm | 1.0× | Cooling is still the largest block, but the part is almost all skin and the window is narrow |
| 1.5 mm | 2.3× | The normal case for consumer housings; packing time now matters too |
| 2.0 mm | 4.0× | Cooling efficiency starts to depend on how well the core is cooled, not just the cavity |
| 2.5 mm | 6.3× | The core is now the bottleneck in most tools; a poorly cooled core doubles the penalty |
| 3.0 mm | 9.0× | Two thirds or more of the cycle is cooling; every point of heat transfer coefficient shows up in output |
The last column is the part people miss. As the wall gets thicker, the cavity side stops being the constraint, because the cavity is a big block of steel with water in it and heat escapes easily. The core is different: it is surrounded by hot plastic, it is often narrow, and there is nowhere to put a decent water circuit inside it. A 3 mm wall on a deep core is not nine times the cooling time of 1 mm — it is worse than that, because the core cannot keep up.
So mold cooling design has exactly one job: get heat out of the steel as fast as the plastic will give it up, evenly, everywhere. There are only three variables available, and they are worth naming before the rest of this page.
- The water temperature. The difference between water and steel drives the transfer rate. It is the variable most people reach for, and the one with the most side effects.
- The flow rate and its state. Turbulent water carries heat away several times faster than smooth water at the same temperature. This is the variable that is usually wrong, and it costs almost nothing to fix.
- The geometry. How far the heat has to travel through steel, and how much surface the water touches. This is fixed once the plates are drilled, unless the insert is printed.
Turbulence is not an option, it is the mechanism
Inside a cooling channel, water can flow in two very different ways. In laminar flow, the water moves in smooth layers and a film of almost stationary water sits against the channel wall. That film is the bottleneck: heat has to cross it by conduction alone, and conduction through stagnant water is poor. In turbulent flow, the water mixes across the whole bore, the film is thin and constantly broken, and the same channel at the same temperature removes several times more heat.
The transition happens around a Reynolds number of 4000. In practice, a mold cooling circuit should be designed for 5000 and preferably higher, which for the usual bores means a water velocity of roughly 1.5 to 3 metres per second.
This is where most tools fail quietly. A mold arrives with eight circuits in series, a 10 mm supply line and a chiller set to 8 °C. The flow is slow, the water is laminar, and the molder compensates by dropping the temperature further — which does not fix the transfer rate, but does introduce condensation on the cavity, higher internal stress in the part, and dimensions that move with the weather.
The cheapest diagnostic in the shop. Read the temperature of the water going in and the water coming out of each circuit. A well-designed circuit runs a difference of 2 to 5 °C. If a circuit shows 10 °C or more, it is not transferring heat properly — either the flow is too low, or that circuit passes through a region that is generating far more heat than the others. Either way, the answer is flow, not a colder chiller.
Bore and flow: the numbers that actually matter
Flow requirements are usually quoted as a total for the mold, which hides the real question. What matters is the flow through each circuit, because that is what determines whether the water in that channel is turbulent.
| Bore | Flow for turbulence (Re ≈ 5000) | Working flow for real cooling (≈ 2 m/s) | Where this bore is used |
|---|---|---|---|
| 6 mm | ≈ 1.4 L/min | ≈ 3.4 L/min | Small inserts, thin cores, bubbler tubes, slides |
| 8 mm | ≈ 1.9 L/min | ≈ 6.0 L/min | The default for small and medium tools |
| 10 mm | ≈ 2.4 L/min | ≈ 9.4 L/min | Medium tools, larger cavities, ganged circuits |
| 12 mm | ≈ 2.8 L/min | ≈ 13.6 L/min | Large tools, cavity plates, high-flow manifolds |
| 16 mm | ≈ 3.2 L/min | ≈ 24.1 L/min | Main distribution lines only — rarely in the mold itself |
The right-hand column is the number to design to, not the middle one. Reaching the turbulence threshold is the minimum for the channel to work at all; running at around 2 m/s is where the transfer coefficient is high enough that further increases stop paying for themselves.
There is a second constraint that decides the bore in practice: pressure drop. A long series circuit of small channels will not reach the flow its Reynolds number requires, because the pump cannot push that much water through that much restriction. A typical mold-temperature controller delivers a few bar at a modest flow, and every extra metre of 8 mm hose, every elbow and every quick coupling takes some of it. This is the reason a circuit that measures perfectly on paper performs badly in the press — and the reason grouped circuits with a manifold usually beat one long series run.
Two practical rules follow. First, size the manifold and the supply hoses larger than the circuits they feed, so the restriction is in the mold and not in the plumbing. Second, put a flow indicator on every circuit, because a circuit with no flow looks exactly like a circuit with flow until someone measures it.
Distance and spacing: where the heat has to travel
Once the water is moving properly, the next limit is geometry. Heat has to conduct through steel to reach a channel, and steel is a mediocre conductor compared with copper — around 30 W/m·K for P20. Two dimensions decide how much of the cycle that conduction costs.
| Dimension | Rule | If it is smaller | If it is larger |
|---|---|---|---|
| Channel center to cavity surface | 1.5–2.0× the bore (12–16 mm for an 8 mm circuit) | Local over-cooling: weak weld lines, high internal stress, and a real risk of cracking or leaking between channels | Heat has further to travel: slower cycle, bigger temperature gradient through the wall, more warpage |
| Channel to channel, center to center | 2.5–3.0× the bore | A cold strip down the middle of the part; visible flow hesitation between channels | Warm bands between channels, uneven shrinkage, sink along the gaps |
| Channel to the parting line | At least 1.5× the bore, and never under 10 mm | Shut-off area too thin: the plate deflects under injection pressure and flashes | Slow cooling at the edge of the part, where the cosmetic surface usually is |
| Channel to ejector pins, screws, inserts | At least 1.0× the bore of clear steel | Drilling breaks into a pin hole, or the plate cracks at the screw; both are scrap | Dead zones where heat accumulates and the part sticks |
| Channel to the outside of the plate | At least 1.5× the bore | Thin wall, water weeps through the side, and the plate cannot be re-machined | No real penalty, just wasted plate |
| Deep cavity or core | Add a circuit every 30–40 mm of depth | The bottom of a deep draw is the last place to cool and the first place to warp | Over-cooling near the top, uneven release |
The trade in the first row is the one designers argue about. Closer channels cool faster and cool more evenly, but a channel too close to the cavity surface creates a cold spot that the part reads as a flow mark, weakens the weld line that passes over it, and leaves only a thin layer of steel carrying the cavity pressure. One and a half bores is the conventional compromise; below one bore, the risk of the plate cracking or the channels leaking into each other stops being theoretical.
The second row is the one that is most often wrong in practice, because channels are usually laid out to a convenient pitch rather than to the bore. Channels 15 mm apart with an 8 mm bore produce a part that is cold where the water runs and warm between the channels, which shows up as visible bands on a glossy surface and as uneven shrinkage underneath it.
Two more habits are worth building in while the drawing is still moving:
- Follow the part, not the plate. A circuit that runs in straight lines across a curved part is at the right distance in some places and too far away in others. The circuit should trace the outline, even if that means more drilling and a few more plugs.
- Keep the cavity and the core on separate circuits. They usually need different temperatures, and they always need different flows. Joining them into one loop guarantees that one of them is wrong, and it removes the ability to correct a warped part by adjusting one side.
Series, parallel, and the imbalance nobody measures
How the channels are connected decides whether each of them gets the flow its bore requires. It is also the part of the design that is easiest to draw quickly and hardest to diagnose afterwards.
| Configuration | Strength | Weakness | Where it fits |
|---|---|---|---|
| One long series circuit | Guaranteed flow everywhere, few fittings, cheapest to build | Water warms as it travels: 5–10 °C from the inlet circuit to the last one, so the two ends of the part cool differently | Small tools, one or two cavities, parts where a temperature difference of a few degrees does not show |
| Parallel circuits from one manifold | Every branch sees roughly the same water temperature | Water takes the easiest path: the branch with the least resistance gets the flow, and the others stall almost completely | Symmetric layouts, and only where each branch has its own flow indicator |
| Grouped series (two or three groups) | The practical compromise: controlled rise per group, balanced flow, still cheap | Needs a manifold and a little more design work | Most medium and multi-cavity tools — the default worth asking for |
| Independent circuits per zone | Lowest temperature spread, and each zone can be set to its own temperature | More fittings, more leak points, more to maintain, higher cost | Precision parts, large tools, hot runner molds, anything where cavity and core need different temperatures |
Parallel circuits deserve a specific warning, because they look like the better engineering answer and frequently are not. Water does not divide evenly between branches — it divides inversely with resistance, and resistance depends on length, bends, fittings and how much scale is already in the channel. A four-branch manifold with one branch 200 mm shorter than the others will starve the long ones, and the tool will cool unevenly in a way that no amount of temperature adjustment can fix. Parallel is only a real choice when the branches are geometrically similar and each one can be checked individually.
For most tools, two or three grouped series circuits are the honest answer: enough flow per circuit, a controlled temperature rise within each group, and a manifold that makes the whole thing visible and adjustable.
The core is the other half of the problem
The cavity side of a mold is easy to cool. It is a large steel block with room for generous channels and direct access from outside. The core side is the opposite: it is the part of the tool the plastic shrinks onto, it is surrounded by hot material on all sides, and its interior is often narrow, deep, and occupied by ejector mechanics.
This is why a tool with excellent cavity cooling and crude core cooling runs at the cycle of its core. It is also why the core is where the money is, whether the answer is a baffle, a bubbler, a spiral core or a printed insert.
| Method | How it works | Best for | Cooling effectiveness |
|---|---|---|---|
| Straight drilled channel | A through-hole, water in one end and out the other | Shallow cores, large core blocks, cavity plates | Baseline |
| Baffle | A plate divides a blind hole; water runs down one side and back up the other | Deep cores with a bore over about 12 mm | Roughly 60–80% of a full channel, falling as the core gets longer |
| Bubbler | A thin tube reaches the bottom of a blind hole and water sprays out at the tip, returning up the annulus | Thin cores under about 25 mm, small bosses, long slender pins | Roughly 40–70%; the only option when the core is too small to baffle |
| Spiral or helix core | A helical groove inside the core forces the water to swirl along its length | High-volume tools where the core sets the cycle | Roughly 90–110% of a straight channel, and much more even along the length |
| Beryllium copper insert | No water at all: a copper alloy insert conducts heat from an unreachable spot to a nearby circuit | Thin standing cores, narrow blades, small features that cannot be drilled | Not comparable directly — but four to eight times the conductivity of mold steel does real work |
| Heat pipe | Phase-change transfer inside a sealed tube | Extreme length-to-diameter ratios where nothing else fits | Rare in production tools; more common in prototypes |
Two things are worth knowing about this table. The first is that the effectiveness figures are not efficiency in the abstract — they are the fraction of a full drilled channel’s heat transfer that the method actually delivers in the same space. A bubbler at 50% still beats no cooling at all by an enormous margin, which is why bubblers are used even on tools where the cycle is not critical.
The second is that the choice is usually forced by diameter rather than by preference. A core under 12 mm cannot be baffled. A core under 8 mm can barely be bubbled. Below that, the answer is either a copper alloy insert conducting heat to a channel somewhere else, or accepting that the core will set the cycle and pricing the tool accordingly.
Aluminum tooling is worth a footnote here, because it changes this arithmetic in one direction only. Aluminum conducts heat roughly four to five times better than mold steel, so an aluminum core cools more evenly and more quickly than a steel one of the same geometry — which is one of the reasons low volume production in aluminum tooling can reach surprisingly competitive cycle times. What it does not fix is the thin standing core: a narrow feature with no room for water is still a narrow feature with no room for water, whatever the material around it.
Hot spots, and how to cool a place you cannot drill
Heat generation in a molding tool is not uniform, and cooling design that treats it as uniform will be wrong in exactly the places that matter. Four locations run hotter than their surroundings on almost every part:
- The gate area, where the melt arrives at its hottest and where the flow is fastest. It deserves its own circuit, or at least its own position on a manifold, because it is the last place to freeze.
- Rib roots and boss junctions, where two sections meet and the local mass is two or three times the nominal wall. These are the classic sink-mark locations and also the classic hot spots.
- The inside of a deep core or a tall standing feature, where there is no room for water and no path for heat.
- Slides and lifters, which are surrounded by hot plastic on several faces, have little steel to conduct through, and are usually not cooled at all. A slide that runs hot seizes, and the seizure is often blamed on lubrication.
For a hot spot that cannot be reached by a drilled circuit, three answers exist. A beryllium copper insert conducts heat out to a channel that is 30 mm away, at four to eight times the conductivity of mold steel. A thermal pin does the same job with a sealed phase-change device. And a printed insert can put a channel where drilling cannot go — which is the subject of the next section.
When conformal cooling is actually worth it
Conformal cooling means a channel that follows the shape of the part instead of being drilled in a straight line. It is made by metal powder bed fusion — printing the insert in tool steel, then machining and polishing it like any other insert. It works, it is proven in production, and it is frequently specified for the wrong reason.
| Aspect | Conventional drilled | Conformal, printed insert |
|---|---|---|
| Distance to the surface | Fixed, and limited to straight lines; corners and deep draws are always further away than intended | Can follow the surface at a near-constant distance, including around corners and inside deep cores |
| Cycle time | Baseline | Typically 10–30% shorter on parts with a dominant hot spot; much less on a simple flat part |
| Warpage | Set by how uneven the cooling is | Better, because the unevenness is what conformal cooling removes |
| How it is made | Gun drilling and conventional toolroom work; predictable, repairable | Metal powder bed fusion, then CNC finishing, polishing and fitting — a different supply chain |
| Material | P20, 718H, S136, H13 forged block | Maraging tool steel powder (1.2709 and similar), or a proprietary tool steel powder |
| Lead time | Baseline | Two to four weeks longer, mostly in printing and post-processing |
| Repair | Weld, re-machine, polish | Welding a printed insert is difficult; a damaged internal channel may be unreachable |
| When it is worth it | Almost every tool, and always the sensible default | Deep cores, thin walls, high volume, one dominant hot spot, or a geometry where drilling cannot get close |
The economic case is narrower than the marketing suggests. Conformal cooling pays when a single dominant hot spot is setting the cycle, when the geometry prevents a drilled channel from getting close to the surface, or when the production volume is high enough that a 15% cycle reduction repays the insert many times over. It does not pay on a flat cosmetic panel with uniform wall, where conventional cooling is already even and the printed insert adds cost, lead time and a repair problem without changing the cycle.
It also pays less often than people expect because the bottleneck is frequently somewhere else. If the tool’s core cannot be cooled and the printed insert is in the cavity, the cycle does not move. Before specifying conformal cooling, the honest question is not “would this be better” but “which specific region is limiting the cycle, and can a printed insert reach it”. If the answer is not a specific region, the money is better spent on a spiral core or on more circuits.
Is your cycle limited by the tool or by the process?
Send us the part drawing and the cycle time you are trying to reach. We will tell you where the heat is actually going — whether the constraint is the core, the circuit layout, the flow rate, or the wall — and what it would take to move it.
Mold temperature is a specification, not a setting
Mold temperature is often treated as a machine setting that the molder adjusts until the part looks right. It is not. It is a property of the resin and the part, and it belongs in the tool specification, because it determines shrinkage, gloss, weld-line strength and internal stress — four things that cannot be inspected out of a finished part.
| Resin | Mold temperature range | Typical setting | What it controls |
|---|---|---|---|
| ABS | 40–80 °C | 60 °C | Gloss and weld-line strength; higher temperature hides weld lines |
| PP | 20–60 °C | 40 °C | Shrinkage and surface gloss; low temperature gives a matte, rougher surface |
| HDPE | 20–50 °C | 40 °C | Warpage, mostly; HDPE is forgiving on temperature and unforgiving on cooling balance |
| GPPS | 20–60 °C | 40 °C | Residual stress and stress whitening on ejection |
| PC | 80–110 °C | 95 °C | Internal stress and crazing; below about 80 °C the part is born stressed |
| PC/ABS | 60–90 °C | 75 °C | Gloss, weld-line appearance, and how easily it ejects |
| PA6 / PA66 | 60–90 °C | 80 °C | Crystallinity, and therefore dimensions after moisture uptake |
| POM | 80–100 °C | 90 °C | Crystallinity and post-molding shrinkage; too cold and it keeps moving for weeks |
| PBT | 60–90 °C | 75 °C | Crystallinity and warp; a hot tool reduces the difference between flow and cross-flow |
| PMMA | 60–90 °C | 75 °C | Optical clarity and flow marks; a cold tool leaves permanent flow lines |
| Rigid PVC | 30–60 °C | 45 °C | Corrosion risk and degradation; both go up with temperature |
| TPU / TPE | 20–40 °C | 30 °C | Release; soft resins grip a warm core and tear on ejection |
| PET (preform or sheet) | 100–140 °C | 130 °C | Crystallinity, appearance and barrier performance; needs a hot-oil or pressurised-water unit |
| PPS | 130–150 °C | 140 °C | Crystallinity and dimensional stability at temperature |
Two consequences follow from that table. The first is that high-temperature resins need equipment, not just a setting: PC at 95 °C, PET at 130 °C and PPS at 140 °C cannot be held by a standard water chiller, and need a hot-water or oil unit with pressurised circuits. A tool specified for those resins but plumbed for 8 mm quick-couplings and a chiller will not reach its numbers.
The second is that mold temperature does not act on the cycle alone — it acts on the dimensions. A tool running 15 °C colder than the resin wants will produce parts that are consistently smaller than the shrinkage figure predicted, and a tool running hot will produce parts that are larger and less stable. This is one of the reasons tight tolerances are so much harder than they look: the temperature that has to be held to make the number is a property of the cooling circuit, not of the machine.
The second is that the two halves of a tool often need different temperatures. A part that warps toward the core can frequently be straightened by running the core 10 °C hotter, which slows the core cooling slightly but changes the shrinkage balance. That is only possible if the cavity and core are on separate circuits with separate controllers — a design decision made before the plates are drilled, and impossible to retrofit cheaply.
Four failures that look like process problems
Scale and corrosion
A cooling channel is a wet steel tube in a warm environment, and it will accumulate scale. A 0.2 mm layer of scale on the wall of an 8 mm channel does not reduce the bore by 2.5% — it adds a layer of insulation exactly where the heat has to cross, and heat transfer can drop by 20 to 40%. The symptom is a tool that ran at 22 seconds for two years and now runs at 27, with no change to the part or the settings. Filters, treated water and an annual descale are cheaper than the alternative, which is a mold whose cycle quietly drifts and whose cycle time nobody can explain.
Imbalance between circuits
Water takes the path of least resistance. In a parallel manifold, one circuit can carry most of the flow while another is nearly stalled, and the mold will show a temperature difference that looks like a design error but is a plumbing error. The fix is a flow indicator per circuit and a valve to trim each one.
Air locks
An air pocket in a high point of a circuit blocks water completely in that section, and the section runs hot while every gauge reads normal. Circuits should be arranged to fill and vent, and the manifold should have a bleed at the highest point. This is a common reason a newly built tool cools unevenly on the first trial and cools evenly on the second, after someone has bled the lines.
No instrumentation at all
The most common failure is not a fault, it is the absence of any means to detect one. A tool with no flow indicators, no pressure gauges and no thermometers on the returns can only be diagnosed by taking the part out and looking at it — which conflates cooling with fill, pack and ejection, and usually leads to a process change that treats a symptom.
Ten questions to ask about the water circuit
- How many circuits are there, and what is the flow through each one? Not the total — the per-circuit figure. If it is not known, it is not designed.
- Is the flow in each circuit turbulent? For an 8 mm bore you want around 6 L/min; for 10 mm, around 9.4.
- What is the temperature rise across each circuit? Above 5 °C is a signal, above 10 °C is a fault.
- How far is the nearest channel from the cavity surface, in bores? Between 1.5 and 2.0 is the target, and it should be stated on the drawing, not left to the driller.
- Are the cavity and the core on separate circuits with separate controllers?
- Does the gate area have its own circuit? It is the hottest region of the part and the last to freeze.
- How is the core cooled — baffle, bubbler, spiral, or not at all? If the answer is “not at all”, the cycle is set by the core regardless of what the cavity does.
- Do the slides have water? Slides that are not cooled seize, and the seizure is usually blamed on grease.
- Is there a flow indicator and a valve on every circuit? Without them, imbalance cannot be found or corrected.
- What is the water quality, and when was the last descale? Scale is the failure that arrives slowly and is never suspected.
None of these questions requires a simulation or a specialist report. They require a mold drawing, a manifold, and someone willing to open the valves and read a gauge. The answers usually explain a cycle that the quotation promised and the press never delivered.
Frequently asked questions
Should cooling channels be 8 mm or 12 mm?
It depends on the size of the part and the geometry of the steel, not on a rule of thumb. An 8 mm channel is right for most small and medium tools: it can be drilled close to the surface without weakening the plate, and it reaches turbulence at a flow a normal temperature controller can deliver. A 12 mm channel is better for large cavity plates and for main distribution lines, where more flow is available and more heat has to move. The mistake to avoid is using small channels for a long series circuit and then finding that the pressure drop will not allow the flow that the bore needs — in that case the answer is a larger bore or more parallel groups, not a bigger pump.
Is a colder chiller better for the cycle?
Up to a point, and that point arrives earlier than most people expect. Lowering water temperature does increase the driving force for heat transfer, and it does shorten the cycle a little. But it also drops the cavity surface below the dew point, which puts condensation on the tool and water marks on the part; it increases internal stress because the skin freezes faster than the core can shrink; and on semi-crystalline resins it reduces crystallinity, which changes dimensions and mechanical properties. In most cases the better move is to keep the water where the resin wants it and raise the flow, which improves transfer without any of those side effects.
When is conformal cooling worth the extra cost?
When there is one clearly identifiable region that limits the cycle and a drilled channel cannot get within about two bores of it — a deep core, a tall standing feature, a curved surface with a hot spot on the inside. In that situation a printed insert commonly takes 10 to 30% off the cycle and improves flatness at the same time. On a flat part with even wall, conventional drilled cooling is already uniform, and the printed insert buys almost nothing while adding two to four weeks and a repair problem. The test is whether you can name the specific region and measure the specific gain, not whether the technology is impressive.
Why does the new mold run slower than the quoted cycle?
Almost always because the quoted cycle was calculated from wall thickness and cooling time on paper, while the tool is limited by something the calculation did not include: a core with no water in it, a long series circuit that cannot reach turbulent flow, a hot spot at the gate, or a slide that runs hot. The quoted number assumes the heat leaves through the steel at a rate the geometry actually allows. When the tool arrives and the rate is lower, the cycle moves, and it moves to whatever the slowest region allows — which is why a fixed cycle time in a purchase order is only credible when the cooling layout has been reviewed against it.
How often should a mold’s water circuits be cleaned?
Annually for a tool in regular production, and more often if the water is not treated or if the tool runs hot. The check is simpler than the schedule: measure the flow through each circuit and compare it with the figure recorded when the tool was new. A drop of 20% means scale, and the fix is a descale before it becomes a cycle-time problem. Filters on the supply, closed-loop treated water and a drain-down before storage all extend the interval, and none of them costs anything close to the production lost to a tool that has quietly slowed down.
Is a mold temperature controller the same thing as the cooling circuit?
No, and the confusion causes real problems. The circuit is the path the water takes through the mold. The controller is the unit that holds the water at a set temperature and delivers it at a pressure. A perfectly designed circuit with an undersized controller behaves like a badly designed one, because the unit cannot deliver the flow the circuit needs, and the temperature it reports is measured at the unit rather than at the mold. When a tool is specified, the flow and pressure requirement of the circuits belongs in the specification alongside the temperature — otherwise the tool arrives and the available unit turns out to be the limiting component.
Cooling design sits at the intersection of three things this site covers separately: how the part’s wall thickness sets the cycle, which is part of the DFM checklist; how the steel conducts and corrodes, covered in mold steel selection; and how the melt arrives in the first place, in hot runner versus cold runner. What the finished tool is proven against, and how the cooling balance shows up in the first samples, is set out in mold trial approval; keeping the circuits clean over the life of the tool is part of mold maintenance; and the whole sequence from drawing to finished part is in the plastic injection molding guide.
