The old fashioned way but still being used by some mold manufacturers. Designing a mold cavity and core in 2D. Some mold designers may use different procedure. You can use Autocad, UGNX, ProE, or just plain tracing paper and pencil in this mold design tutorial.
1) Calculate for the mold dimension using the shrinkage factor.

The shrinkage factor can be determined by resin material properties or by experimenting. For example the PBT has a shrinkage of 18/1000.
Compensate for the tolerance and other possible deformation.
Include the draft angle whenever possible. The draft angle should be within the dimensional tolerance.

2) Draw product drawing using the calculated mold dimension. Include the embossed texts if it is a part of the product drawing. It is a good idea to draw the embossed texts using lines and curves. If your CAD is capable of “reflecting a text” then you are in advantage.

3) Mirror or flip your product drawing. The mirrored drawing will be your mold drawing. Notice that the embossed texts were also mirrored and they became engraved text.

4) Decide and draw the gate location. Locate it away from small core pins to avoid damaging those pins during resin injection process. For our example, I would like to use “side-gate”.

5) Decide the parting line. Input the parting line changes if there are any. Parting line changes should be visible on the top view, draw that too.

6) Decide the ejector location. Divide the mold drawing as you wish or as your process capability would dictate. Consider dividing on gas vents.

7) You can then derive or trace your cavity and core drawings using your mold drawing as reference.

The current landscape of injection molding heavily relies on classical simulation methods, primarily 2D and 2.5D modeling techniques. While these methods have historically served the industry well, they have significant limitations that can impede efficiency and innovation, particularly as product designs and materials grow increasingly complex.

1. Limited Dimensional Accuracy
Classical simulations often fail to capture the full three-dimensional complexities of mold designs and the flow of the molten materials within them. This limitation stems from the inherent simplification in 2D and 2.5D approaches, where important effects such as the interaction between different material layers and the precise behavior at corners and undercuts are either approximated or completely ignored. These oversights can lead to inaccuracies in predicting defects, stress points, and the final strength of molded parts.

2. Inadequate Representation of Flow Dynamics
Traditional methods typically handle the flow of polymers as either ideal or simply non-ideal fluids without fully accounting for the unique characteristics of injection molding materials, such as their non-Newtonian behaviour. This results in a poor prediction of how materials will fill the mold, affecting gate location, pressure, and overall cycle time decisions. Additionally, the cooling phase, crucial for determining the product’s final properties, needs to be more adequately modeled, leading to issues with warpage and residual stresses.

3. Slow Iterative Processes
Classical simulation tools usually operate on a slower iterative basis, where each change in design or processing parameters requires a new simulation run. This can be time-consuming and inefficient, particularly in a development environment where multiple iterations are the norm. The lack of real-time feedback delays decision-making and extends the product development cycle.

4. Scalability and Resource Intensity
Many traditional simulation programs are resource-intensive, requiring substantial computational power and time, particularly for larger or more complex models. This limits their scalability and accessibility, especially for smaller manufacturers or those dealing with highly intricate products. As a result, some companies may opt out of thorough simulation testing altogether, which can compromise product quality.

5. Generalisation Over Customisation
Finally, classical methods often generalise material properties and processing conditions rather than offering the ability to customise parameters based on specific real-world conditions. This generalisation can lead to less optimised manufacturing processes and a greater likelihood of product failure due to discrepancies between the simulated and actual conditions.