Warpage, cooling and an inspired guess

Shaving the seconds necessary to get cycle times in line with productivity targets involves some compromises, and the best place to start is with the fattest part of the cycle, cooling time.

“Cooling” always seemed a misnomer to me, since most processes in which I was involved, parts were ejected hot, damn hot, and burns on the operator’s fingers (ejection was an imperfect art in our shop) were avoided with layers of green “grip tape”.

Cooling just enough to get the part solid, then opening the mold is an old technique, but what happens when the new, faster machine timing results in warped parts? It’s common, and happens as a result of differential cooling rates on different areas of the mold cavity. Differential cooling? If you’re thinking that cooling is supposed to be balanced, you’re right, but consider this: is it balanced based on the old, longer cooling time?

In many cases, imperfect balance can be compensated for by raising the mold temperature and keeping it shut longer to minimize the cooling rate differential across the cavities. This isn’t a problem if mold and machine are running at design specs, but alter (or maybe “optimize”) even simple parameters and things can go horribly wrong in a hurry. Salvaging efficient mold cooling in these cases is still possible, however. I would usually start by thinking about what caused the problem in the first place. Frequently it is “simply” unsymmetrical distribution of cooling lines, usually due to a core pull, weird ejector design or because of the part configuration. Since you can’t reconfigure the jacket, modifying how the water flows is a logical approach.

The temptation is to play with lowering temperature immediately, but in my experience a higher “delta T” just makes the situation worse. It can be a big problem with flat parts, or parts that need complex and difficult-to-cool cores. A trick that has worked for me on occasion is to reverse the flow in the mold, and use flow restrictor bushings in the lines. I suspect that the flow restriction is the key, for two reasons. Slowing flow in a specific parallel branch of a cooling circuit (don’t try this with a series circuit, believe me) raises the temperature of the cooling water in the branch, locally slowing the rate of heat transfer out of the mold. That’s because the rate that the heat flows is proportional to the temperature difference between the mold steel and water.

My other theory is that restriction can induce turbulent flow in the system, which promotes more even cooling than “smooth” systems with areas of less efficient (for cooling) laminar flow. The trouble is, Band-Aid solutions like this operate by trial and error, or an inspired guess, so it’s important to make the restriction easy to change. My “inspired guess” was to replace the plug on one of the mold’s cross-drilled “galleries” with a longer set-screw, ground down to create a projection that intersects the corner of the cooling channel. By grinding several on a surface grinder with different projection sizes (“projection” wasn’t the word used in our shop, of course) it’s relatively easy to meter flow in one or more branches as necessary.

The important point to remember if you want to experiment with this low-cost technique is this: don’t change other parameters until you’ve explored the restrictor plug option completely. If you don’t, there’s no way to know which change worked, which didn’t, and which might have if you hadn’t fiddled with every control on the chiller.