A New Look at Evaluating Fill Times For Injection Molding
One of the important process parameters to establish and record for any injection molded part is its injection or fill time. Research research reveals the limitations of popularly taught methods of establishing this critical parameter.
John Beaumont, Beaumont Technologies Inc.
Injection molding process methodologies have evolved over the decades from a seat-of-the-pants black art to a more structured approach. A number of schools, companies, and individuals provide a valuable service to the industry by teaching these structured methods, which have been labeled with terms such as Scientific Molding, Decoupled Molding, and 2-Stage and 3-Stage processes. These approaches involve similar specific procedures that help establish a foundation on which to build a process. Among the procedures taught are a method for determining a fill time or fill speed using in-mold rheology curves (a.k.a “relative viscosity vs. relative shear rate” curves), as well as methods for establishing an ideal transfer position, ensuring the proper melt temperature, finding the ideal hold pressure, identifying pressure losses within the mold, and finding the time when gate seal (gate freeze) occurs.
These approaches also teach that the process must be documented in a manner that allows it to be transferred to other molding machines with the intent of achieving relatively consistent part quality. This requires that the process is recorded by referencing plastic variables, not machine variables, and is done using a “universal setup sheet” (a term used by John Bozzelli, scientificmolding.com). For example, if you are documenting the melt temperature you would document the temperature of the plastic coming out of the machine nozzle—not the barrel temperature settings on the machine controller.
ESTABLISHING OPTIMUM FILL TIME
One of the important process parameters to establish and record for any injection molded part is its injection or fill time. Fill time is an indication of how fast the plastic is injected into the mold. Fill time affects how much shear heating and shear thinning the plastic experiences, which in turn affect the material’s viscosity, the pressure and temperature of the plastic inside the cavities, and the overall part quality (dimensions, aesthetics, strength, etc.). Any change in fill time may adversely affect the final molded part. Therefore once the ideal fill time is established for a given mold, that fill time should live with the mold forever and should be allowed to vary only slightly (±0.04 sec, as per John Bozzelli’s recommendation). The key question is: How does one go about identifying an ideal fill time for a given mold?
Molders use several methods to establish a fill time, some of which begin with one of the following methods:
•Evaluating the fill time used on similar parts and molds.
•Trial and error.
•Design of experiments (DOE) data.
•Relative Viscosity test.
Ideally, every part would be evaluated for fill time using mold-filling simulation performed by a skilled analyst with plastics processing experience. Unfortunately this type of analysis data is not available for many plastic parts, so molders need a method to establish an ideal fill time that they can employ on the shop floor. This is where the Relative Viscosity (RV) test comes into play. The general procedure for this commonly taught method is presented below.
Though some consultants and trainers may add other steps or teach it a bit differently, essentially the approaches are very similar:
1. Using maximum injection velocity, adjust shot size to get the fullest part 95% full.
2. Record the fill time and pressure at transfer.
3. Reduce injection velocity and record the fill time and pressure at transfer.
4. Repeat Step 3 until the fill time is over 10 sec.
5. Use the data to calculate relative shear rates and relative viscosities:
Relative Shear Rate = 1 ÷ fill time
Relative Viscosity = Hydraulic pressure x intensification
ratio x fill time
6. Graph relative viscosity (y-axis) vs. relative shear rate (x-axis).
7. Select a fill time on the “flat” portion of the curve.
The results of one RV test are shown graphically in Fig. 1. Note that the graph is essentially that of “Fill Time vs. 1 ÷ Fill Time” (where the Fill Time on the y-axis is multiplied by its
corresponding pressure). When you graph a number versus its reciprocal, the shape of the curve shown in Fig. 1 is expected. If the change in pressure were directly proportional to the change in fill time, the graph would show a straight line. However, this test demonstrates that as plastic flows faster (higher shear rate), its viscosity is reduced and becomes fairly consistent over a wide range of injection rates.
The RV test results in “viscosity vs. shear rate” curves that look similar to those produced using laboratory capillary rheometers (Figs. 2A, 2B). Figure 2A shows the capillary rheometer data of viscosity (poise) versus shear rate (1/sec) plotted on a log-log graph while Fig. 2B shows the same data plotted on a non-log-log graph. The non-log-log graph in Fig. 2B looks very similar to the RV graph in Fig. 1.
The RV test has been widely taught and adopted by many molding companies to help select a fill time for a given mold. Company procedures have been written to include the test as part of their process standards. Thus, there is a great deal of material, labor, and machine time spent each day on running the RV tests in hopes of identifying optimum fill times based on a structured approach that a process technician or engineer can use on the shop floor.
Ideally, someone applying this technique should also be evaluating the parts for cosmetic issues and other defects to help determine the optimum fill time. But all too often we have found molders looking at the curves as a scientifically founded method to establish optimum fill time and therefore assuming they should look at other process parameters to address molded part problems.
Those who teach the RV test method usually state that for most parts the fill time should be located somewhere on the flatter portion of the curve (after the “elbow”). The flat portion
of the curve is considered desirable because it is expected to produce the most consistent melt conditions when changes occur in the process or resin viscosity. These changes might result from lot-to-lot material variations or process drift.
Depending on the range of fill times used, the flat portion of the curve can be a small area or rather large. This leaves the processor asking, “Where on the flat portion of the RV curve is the optimum fill time?” Currently the practice of picking the fill time from the curve is quite arbitrary. If you ask three different molders taught in the same class to select a fill time from the same RV graph you may get three different answers. Listed below are several common opinions and one formula that have been suggested as guidelines for selecting the target fill time:
Method 1: Halfway from the “knee” to the end of the curve.
Method 2: The point right after the “knee.”
Method 3: The farthest point out, because it has the lowest viscosity overall.
Method 4: ([(Highest RV – Lowest RV) x 0.05] ÷ 2) + Lowest RV.
A further complication is that the selected fill time is completely dependent on the scale of the graph at hand. If the scale of the graph changes (i.e. run the test to only 5 sec instead of 10 or 20 sec), the fill time chosen by using the current visual and arbitrary fill-time selection methods may actually change (Fig. 3A vs. 3B). Also, the ideal fill time may be harder to determine since the flat part of the curve will not look as flat as it does when running the test with more data points at longer fill times.
In addition to some ambiguity in the interpretation of the RV test, the method of focusing on a “relative viscosity” to identify a fill time raises further questions. It was apparent that further research was required to better understand the RV test and determine if it could be improved upon. This research began several years ago and only a portion is reported here.
RESEARCHING THE FILL-TIME PROBLEM
Early stages of this work began with us asking ourselves, “What do relative viscosity and relative shear rate really mean?” The relative shear rate is defined as 1 ÷ fill time, which results in units of 1/sec, the same units used by rheologists for describing shear rates. However actual shear rate depends on the volumetric flow rate, the non-Newtonian characteristics (n), and the specific geometry through which the plastic is flowing, as per the following equation: