Every cooling problem in a mold comes down to one question: how fast can we pull heat out of the steel? In practice we describe this with three boundary conditions. The first kind fixes the temperature on the surface—useful for a quick estimate of a thick core insert sitting on a water-cooled plate. The second kind fixes the heat flux, which is what we actually deal with at the polymer interface: the melt delivers a known heat load (roughly 0.5–2 W/cm² for typical injection molding, higher in thick sections and hot runners) that must be removed each cycle. The third kind is the convective condition, where a surface at temperature Ts meets a coolant at Tf and the heat flux is q = h(Ts − Tf). This is the one that governs the waterline wall, and it is where most mold designers get sloppy.
The convection coefficient h is not a constant you can look up once. For turbulent water flow in a 10 mm cooling channel at 1–2 m/s, h typically lands in the 5,000–15,000 W/m²·K range; drop the velocity and you fall into transition flow, where h collapses and cycle time climbs. That is why baffles, bubblers, and spiral plugs matter—they keep local velocity up in deep cores and tall ribs where a straight drilled line cannot reach. Reynolds number, channel diameter, and coolant temperature all feed into h, so a line that works on a 6 mm channel may underperform badly if you neck it down to 4 mm at a tight feature. Calculate it, don't assume it.
In real tooling, the three conditions rarely act alone. A cavity wall sees a heat-flux condition from the melt on one side and a convective condition from the waterline on the other, with conduction through the steel in between—so the steel's thermal conductivity and the distance from the wall to the waterline centerline (typically 2–3× channel diameter) become design variables, not afterthoughts. Get these numbers right and you can predict cycle time before the first cut. For more mold sourcing and technical guidance, visit MoldSo at www.moldso.com.