Start with the simple fact that plastic resins shrink as they cool. Some shrink more than others, but they all do it. If the shrinkage were perfectly even, we could simply make the mould slightly oversize and count on shrinkage to reduce them to the desired size. Unfortunately, shrinkage is a more complicated process. As a result, certain shapes that are otherwise perfectly acceptable can be difficult to mould because they tend to warp as they cool. Anything with a "C" shape, like the part shown in Figure 1, can be particularly problematic.
Of course your choice of resin can contribute to the problem in two ways. The first is variation in the tendency of the resin to shrink as it cools. For example:
- Acrylic shrinks very little
- HDPE shrinks quite a bit
- Nylon 6/6 falls somewhere between the two
The second materials issue is specific to filled materials. As they are injected into the mould, the fibre filler in these materials tends to align with the direction of resin flow. The resulting "grain" causes uneven shrinkage between dimensions that run with the grain and those running across the grain. The result is an increased tendency of parts made of filled resin to warp as they cool.
As far as shapes are concerned, the problem with "C" is actually a problem with "L." The problem lies in L's right angle corner, and C, having two right angles, is twice the problem.
Figure 2 is a close-up view of the angle of an L. You can see that the distance along the inside of the angle (from A to B) is shorter than the distance around the outside of the angle (from C to D). As a result, the surface on the outside of the angle is larger than that on the inside. More area means faster radiation of heat. As a result, the C-D side of the angle hardens before the A-B side. As A-B continues to cool, it also continues to shrink, pulling what was designed to be a right angle to something less than 90 degrees.
The L doesn't have to be an actual right angle to have this problem. Any curve will be shorter on the inside than on the outside. (See Figure 3) But, for a given wall thickness, the difference between the area of the outside face and the area of the inside face will increase as the radius of the curve decreases. (Think of a sharp right angle as a curve of infinitely small radius.) Also, the sharp edge of the mould stays hot longer, increasing the differential. For these reasons (along with the fact that radiused corners build up less stress) you should always consider the option of radiusing corners if possible.
Of course, everything that's true of an L is doubly true of a C, which increases the magnitude of the problem because there is more curve, hence more difference between the length of the outside and inside surfaces. Whether made up of angles or curves, the inside of the C will be shorter than the outside and, as a result, will still be cooling and shrinking after the outside has hardened, pulling the "jaws" of the C closer together.
There are a number of ways to address the problem. Turning the C into an "O" eliminates the opening and prevents the ends of the C from being pulled toward one another. In essence, the added part of the circle acts as a brace to help the part hold its shape. Putting a removable brace across one of the open sides can also help counteract the forces trying to close the jaws of the "C."
If none of these is possible, the best way to reduce the problem is to choose one of the more "shrink-resistant" resins. These would include: ABS, Polycarbonate, PC/ABS, PETG Polyester, Polystyrene, and K-resin Polystyrene butadiene. And, of course, where shrinkage could distort your part it is particularly important to pay close attention to geometry and avoid filled resins.