Conformal cooling: the case that almost always pays off

Of all the applications of metal 3D printing, conformal cooling in mold inserts is the one that is most frequently cost-effective, and for a reason that has nothing to do with geometry: it is the only one in which the savings are realized on each injected part, thousands of times, instead of just once.

The problem with drilled channels

An injection mold is cooled by circulating water through internal channels. With conventional machining, these channels can only be straight: they are drilled, and the excess is plugged.

This leads to two problems. The channels pass far from the areas of the cavity that are not within their path, so those areas cool by conduction through the steel, much more slowly. And the distance between the channel and the surface varies along the length of the part, so some areas cool faster than others.

The consequences are known to anyone who injects:

The cycle is determined by the slowest point. You can't open the mold until the hottest area has solidified, even if the rest has been ready for a while.
Uneven cooling causes warping. The part shrinks more where it cools later, resulting in a warped shape.
Shrinkage marks appear precisely in the thick, poorly cooled areas.

What changes with a channel that follows the surface

Additive manufacturing allows the channel to follow the geometry of the cavity at a constant distance, between and around areas where a drill bit cannot fit.

This achieves what matters: uniform distance, and therefore uniform heat extraction. The slowest point is no longer so far behind the rest, and the cycle time decreases.

The ranges typically seen in the industry for a well-chosen application are cycle time reductions of around 15 to 40%, with less warping and fewer shrinkage cavities. The range is intentionally wide: it depends entirely on whether your mold has an identified hot spot or not.

Whether it's worth it or not is decided with a calculation, not with a hunch

It's a four-line calculation and it's done before drawing anything.

How long is the current cycle, and what fraction is cooling time? If cooling is 6 out of 30 seconds, the potential for improvement is small. If it's 25 out of 40, there's significant potential for gain.
How many parts per year come out of that mold? Savings are multiplied here. With short production runs, it's not worth it, no matter how good the mold is.
What is the machine cost per second? Hourly cost of the injection molding machine divided by 3,600.
Multiply. Seconds saved × parts per year × cost per second. If the result recoups the additional cost of the insert in less than a year, it's a clear yes.

The second question is the one that rules out the most cases, and it's the one that is almost never asked first.

How to design the channel

Distance to the cavity: constant, typically between one and two times the channel diameter. Too close compromises the insert's mechanical strength; too far negates the effect.
Cross-section: preferably teardrop or elliptical rather than circular, so the channel is self-supporting during fabrication and supports don't need to be inserted. A support inside a closed channel cannot be removed.
Internal roughness: the internal surface of a LPBF-manufactured channel is rougher than that of a drilled one. This increases pressure drop—you have to check that the pump can reach it—and, interestingly, improves heat transfer by promoting turbulence.
Powder evacuation: the channel must be able to be emptied of the unmelted powder that remains inside. If it has crevices where the powder collects, that powder ends up in the cooling circuit.
Watertightness: this is the most tested and critical aspect. An insert with a water leak inside a hot mold is a serious problem.

Material

The usual one is a maraging tool steel type Fe2709 or C300: it is manufactured well by LPBF, hardens by aging with minimal deformation —which avoids having to grind much later— and withstands the thermal cycling of the mold.

When the hotspot is highly localized and a large amount of heat needs to be dissipated from a small area, copper alloys are used, which conduct heat much better at the expense of lower mechanical strength. It's a design decision: copper where heat is the primary factor, steel where the load is the primary factor.

When not

For the sake of honesty, here are the cases in which we do not recommend it:

Short production runs. Without volume, there's no possible return on investment.
Molds without an identified hot spot. If the cycle is dictated by something else, improving cooling makes no difference.
Very large inserts. They exceed the build volume and have to be cut, resulting in seams where they're not needed.
When no one has measured the current cycle. Without baseline data, the improvement can't be demonstrated, and the investment is justified on faith.

—

If you have a head insert and want a price quote, the quote calculator will give you an estimate of the manufacturing cost. Available alloys are listed in the materials catalog, and design rules that avoid supports within the channel are found in supports in metal 3D printing.

Why this specific case is not reached by either the foundry or the MIM, in printing, melting or sintering.