Cooling is the single largest factor in injection molding cycle time, and the one most often under-designed. A tool with poor cooling runs slower, produces parts that vary in shrinkage from cavity to cavity, and warps as the tool warms up. Every one of those problems shows up in the part price long before anyone blames the cooling layout.
Key takeaways
- Cooling decides cycle time. A large share of the cycle on most tools is not filling or packing — it is waiting for the part to reach a temperature the ejector can handle.
- Hole-to-surface distance is the rule that matters. Coolant drilled too far from the cavity face has to travel through steel to do any work.
- Uneven cooling causes three different defects, and they look like different problems until you trace them back to the layout.
- Mold temperature changes the material, not just the cycle. Crystalline resins and amorphous resins respond to different mold temperatures for different reasons.
- Cooling is designed on paper, before the tool exists. Adding circuits after machining means re-drilling inserts, and a re-drilled insert is a new insert.
Why cooling dominates cycle time
An injection cycle has four parts: fill, pack, hold and cool. Fill and pack are measured in fractions of a second for a correctly sized part. The cool phase is not, and it is the phase the machine is idle while the tool does its most important work.
That is why the fastest tools are not the ones that fill fastest. They are the ones where coolant reaches the cavity face quickly and evenly, so the part reaches ejection temperature in the shortest possible time. Halving the time to a safe ejection temperature can cut a 40-second cycle to 25, which is a 37% throughput gain before a single process change.
Our cycle time reduction guide covers the other levers, and the honest caveat is that cooling is usually the biggest one.
Hole-to-surface distance: the rule that decides everything
Coolant only does work where it is. The practical guideline most shops work to is that the distance from the coolant channel to the cavity surface should be roughly equal to the channel diameter, and never more than about 1.5–2 times it.
Drill a 10 mm channel and leave 40 mm of steel between it and the cavity and the coolant is warming before it arrives. The surface cools slowly and unevenly, and the gradient across the cavity face produces exactly the defects buyers complain about later. A channel drilled too close to the face is a different problem: it risks breakthrough into the cavity, and once you break through, that cavity is scrap.
This is the single most valuable thing a DFM review can catch, and it is invisible in a 3D model. Our DFM rules reference covers the other design-level checks worth doing before steel is cut.
Straight, U-shaped and follow-core circuits
Three layouts cover almost everything, and the choice is driven by the cavity surface rather than by preference.
- Straight (bypassed) channels enter and leave through the plate, ideally along a parting line or a side wall that keeps them out of the steel the part sees. Cheapest and fastest to drill, and the standard choice for flat panels and simple ribs.
- U-shaped channels go in one side, turn, and come back out the other, closer to the surface than a straight pass. Used on thicker walls and boss features where straight drilling would sit too deep.
- Follow-the-core circuits are drilled to track the geometry of a core or insert, so coolant follows the heat path rather than the shape. These need inserts, which cost more and take longer, and they are the right answer on cores that run hot enough to distort the part.
The three defects uneven cooling produces
Buyers usually arrive with one of three complaints and rarely connect them to the same cause.
- Differential shrinkage. One area of the part cools faster than another, so different regions shrink at different rates. The part arrives within tolerance on the first shot and out of tolerance on the four hundredth.
- Warpage. Asymmetric cooling across a flat panel makes one side shrink more than the other, and the panel bows. Warp is a cooling defect until proven otherwise.
- Sink marks that never quite disappear. A thick section that cools too slowly keeps drawing material after the gate has sealed, so the surface under the gate stays slightly depressed. More pressure will not fix it; only cooling time or geometry will.
Each of these gets treated as a process problem, adjusted for a few days, and then accepted as normal. That is the expensive path. The injection molding defects guide covers the full defect set, and it is worth reading before adjusting anything.
Mold temperature and what the resin does about it
Cooling is not only about cycle time. Mold temperature changes how the material behaves, and the direction depends on the resin family.
Amorphous resins such as ABS, PC and PMMA shrink predictably as they cool, with no crystallization phase to manage. A warmer mold means a longer cycle but a better-looking, more stress-relieved part.
Semi-crystalline resins such as PA66, PP and POM have a crystallization phase. Crystallization rate rises with mold temperature, so a cold mold produces a part with lower crystalline structure and different shrinkage than the same part molded hot. This is why a PA66 part that met spec on a warm tool can drift out of tolerance on a cold one.
Our PA66 vs POM comparison goes deeper on that family, and the materials guide covers the rest of the range.
Cooling is also a tolerance problem
Shrinkage is not a single number, and a tool that holds one value only at one mold temperature is not holding a tolerance. It is holding a coincidence.
This is where quoting accuracy separately matters. Our machining tolerance is ±0.005 mm on critical mold dimensions. The molded part tolerance that tool can then hold is ±0.01 mm on critical dimensions, subject to the resin and the part geometry. Those are different numbers answering different questions, and a supplier who quotes a single figure for both has not thought the problem through.
On tight-tolerance parts, specify the mold temperature the part is measured at. Without that, a dimension taken hot and a dimension taken cold are not the same measurement. Our tolerances guide covers how to write that into a specification.
Cooling and the tool steel
Steel conducts heat differently from aluminum, and the difference shows up directly in the layout. Aluminum prototype tooling accepts simpler cooling because the metal itself moves heat fast. In hardened steel, the coolant does the work, so the circuit layout has to be designed rather than assumed.
For high-volume production tools our reference is 1.2344 (H13) and 1.2343 (H11) — hot-work steels hardened to 48–52 HRC. Tool life on those is up to 1,000,000 shots, depending on steel grade and maintenance. The steel selection guide covers the trade-off between the common grades.
Machine capacity matters here as well. Cooling circuit quality depends on accurate drilling, and drilling accuracy depends on how well the machine holds position over a long run. We run 13 CNC machining centres with travels up to 1,600 mm, 6 EDM machines including Sodick mirror-finish and twin-head units, and 3 wire EDMs.
Conformal cooling, and when it is worth it
Conformal cooling — printed or sintered channels that follow the three-dimensional surface of the cavity face — is the only way to cool a complex geometry evenly. On a deep, thin-walled, high-cavity part it can cut cooling time by more than half.
It is not the default answer. Printed inserts cost more, lead times are longer, and the channel density leaves less steel behind the cavity face than a drilled core. For a straightforward two-plate tool with flat panels, a drilled circuit is faster to build, easier to repair, and produces the same result. Conformal cooling earns its cost on the geometry that drilled circuits cannot reach, and nowhere else.
How to verify the cooling layout before the steel is cut
The layout should be checked thermally, not judged by eye. At RCH Plastic your STEP or IGES file is modelled in UG, run through Moldflow, and returned as a written DFM report within 3 working days, at no cost. The report covers fill, pack and the thermal behaviour of the proposed layout, and it is the cheapest insurance available in the whole project.
Read the report before the tool is programmed, not after. Every change made after drilling is a change to inserts, and inserts are one of the most expensive line items on a tool. The mold flow simulation guide explains what to look for in the results.
Getting the trial right
Cooling problems do not show up on the first shot. They appear as the tool heats to steady state over a few hundred cycles, which is why trial budget matters. Molding trials cost $300–$800 each, and at least one or two should be budgeted before approval.
What you should receive at trial is not a box of samples. It is samples, trial photos, a trial video and the full process parameter sheet, together with a CMM dimensional report on the critical dimensions. Every claim should be backed by a document issued against your mold number. You approve from measured data, not from a promise.
Cooling problems are tool problems, not machine problems
Buyers frequently try to solve cooling problems by changing the machine settings — higher pressure, longer hold, faster screw recovery. Those changes hide a cold spot briefly and then push the problem somewhere else, usually into flash or breakage.
If a part is warping, shrinking unevenly, or drifting out of tolerance over a shift, the fix belongs in the cooling layout. Adjusting the process to compensate for a bad layout converts a one-time tool cost into a permanent process cost, and it is the more expensive of the two.
What a cooling review should return
When a supplier quotes a tool, ask for these six things. If they are not available, you are buying a guess rather than a design:
- Cooling time and total cycle time, broken out rather than quoted as one number.
- Estimated mold temperature at the cavity face for your run time, not a generic range.
- Part weight and shot weight, since both set how much heat has to be removed.
- Which regions are deliberately hot and why — every tool has at least one, and it is usually intentional.
- The trial plan, including how many cycles before dimensional readings are taken.
- What happens at steady state, not only at first shot.
Tooling and part cost, separated
Cooling design is a tooling cost. It shows up once, in the tool invoice, and it pays for itself through the part price for as long as the tool runs. That distinction is worth keeping clear when a quote looks high: 60–70% of a mold bill is machining hours on the cavity and core, and the cooling layout is a large part of that.
Our injection molding cost guide covers the full pricing picture and the cost breakdown guide itemises where every dollar lands, including the inserts that conformal and follow-core cooling require.
Related guides in this series
- Mold flow simulation guide — how cooling gets verified rather than assumed.
- How to reduce injection molding cycle time — the levers around the cooling phase.
- DFM rules for injection molded parts — the design decisions that make cooling easier.
- How tight can injection molding tolerances really be — why shrinkage is a range, not a number.
- Multi-cavity mold design — how cooling scales when the cavity count goes up.
Get a cooling review with your quote
Send the 3D file, the resin, your cycle time target and your annual quantity. We will return the cooling layout alongside the quotation, with the cycle time broken out — and if your current supplier’s layout is leaving performance on the table, we will tell you where. Request a quote with cooling review.