Mold Cooling: A River Runs Through It

Think cooling an injection mold is as simple as piping cold water through it? I once did, until two events changed my perspective. One was a temperamental 100-ton press that had to crank out six cavities worth of small PP auto parts to feed an assembly line. We were running the press overtime to keep up with the line, and margins didn’t allow much investment in new equipment (sound familiar, OEM suppliers?). So we did what anyone would do: shaved seconds off the cycle time.

When it came to the chill portion of the cycle, however, I hit a brick wall. To get all the cavities to deliver in-tolerance parts, there seemed to be no way to save time while the resin froze. Naturally, I took some graveyard shift downtime and began the experiments. Shortening the feed lines by moving the chiller closer had no effect; neither did dropping the set temperature, which was almost maxed out anyway. I “rodded” the mold, and scrubbed the channels like my Winchester 12-gauge, but couldn’t win a second; I even radiused the hose barb inner “steps” to try to smooth the coolant flow.

Nothing.

What eventually did work was a change to the coolant routing in the mold, but not for the reasons I suspected. I started by reversing the flow by switching the inlet and outlet (something I wouldn’t do unless the tubes were super clean) and began to see a little progress. On a whim, I split the flow into two circuits, feeding from the edges first and then from the centre. Now we saw real savings. Why? The answer only came to me much later, during a heat transfer course at the University of Toronto. I knew from the shop floor that the temperature gradient between inlet and outlet at the mold was important to the cooling cycle, but in the classroom the story was more complex.

Temperature was a parameter you measured, but heat is what you’re trying to remove. Guess what? They’re not the same thing. And the rate of heat removal in a tube heat exchanger (mold) is dependent on the temperature difference between the coolant and the metal wall of the gun-drilled mold channels, among other things. Cavities near the inlet side of the mold had the biggest temperature difference (we called it “Delta T”) and cooled the fastest. It was much slower on the outlet side because the coolant had warmed considerably by the time it reached the other side of the mold.

Reversing the flow direction only swapped the problem from one cavity to the other. Splitting the flow into two circuits, however, shortened the length and reduced the surface area of hot mold felt by cold inlet water because two inlets meant that each half circuit only chilled half the mold. Two circuits also split the mass flow rate of the cooling water, allowing better heat transfer, adding even more efficiency. And the cost was little more than a few quick connects, brass fittings and some feed line.

So what did I learn? Your mold hold time will be determined by the freeze time of the last cavity to cool. Cranking up your chiller or adding capacity will help, but the lead-lag effect will still be there. Multiple, shorter circuits narrow that spread and save precious seconds, all else being equal.

Keep in mind, this strategy worked because the mold was symmetrical; try it with a family mold or odd shape and, as they say in the TV commercials, your results may vary.

It’s good to know your “Delta T” between inlet and outlet, but that’s not the same as knowing the cooling rates at each cavity. I’ll expand on this a little in the next issue of Canadian Plastics.

And by the way, if you decide to play with coolant routing, don’t forget to bring safety glasses, lots of rags and a mop and pail. (Don’t ask me how I know this.)