I received this common question via e-mail the other day…
I heard that it’s good to have shot size from 20 to 80 percent of the machine’s available shot size. Has this conclusion was tested, or has somebody published a white paper study?
Although I am not familiar with studies to validate this specific conclusion, but you can find studies on many of the aspects which contribute to this general rule of thumb. I can give a good argument to support both the 80% and 20% limits. Even more important… is the fact that you can easily generate your own data to validate/test any of these arguments with your specific processes.
80% – This rule of thumb is provided to give a buffer to allow for process variation for a couple compounding reasons… (1) A good cushion should be between 5-10% of the overall shot size. (2) Many machines require 2-5% of the shot size to decompress the screw after recovery. (3) The check ring will typically vary 2-5% during fill resulting in a similar variation in cushion size. When you add these variations up, you need a 10-20% buffer to help ensure you can properly fill the part.
20% – The typical general purpose screw contains approximately 1-2 shots of material within the flights of the screw. This means that a process running at 50% capacity will have an estimated barrel residence time between (2) and (4) * (cycle time). Likewise, a machine running at 20% capacity has an approximate residence time between (5) and (10) * (cycle time). If you bring this to the extreme, a process running at 5% capacity could have a barrel residence between 2000% and 4000% of the cycle time!
These are rules of thumb, and therefor there are always exceptions. With the use of Accurate process controls and short travel check rings, you may be able to violate the 80% rule. I never recommend violating the 20% rule as it is likely to affect the part quality, process stability, and it will waste a large amount of energy.
Unlike most manufacturing processes, the compressibility and shear thinning characteristics of plastics cause inherent variability in the process.