Plastic Injection Troubleshooting:
Using Scientific Molding Approach When Validated Processes Fail
Many companies apply scientific principles to their set up and validation of processes. After all, this methodology has proven its worth as a successful approach in process development. However, it is important to note that SM goes beyond set up. The true test of processors is how they evaluate change in validated processes that have already been proven. This article will define criteria that identify root causes for change and provide information and approaches to evaluate and correct changes that occur.
Three Steps of Process Development and Validation
The key to successful troubleshooting when changes occur is understanding that the initial validated process serves as the foundation, and variances outside of control limits established are detrimental to established process control parameters.
A validated process is defined as 100%+ efficient based on quoted cycle, minimal (< 1%) to zero scrap and has run a minimum of 8 hours at these efficiency/ scrap levels. In addition, the goal for a validated process at start up is packing acceptable parts by the 3rd shot at start up.
It is important to note that these requirements may not always be achievable, but it is a processor’s goal to meet these expectations. In terms of principles of validation, there are 3 steps to process development and validation:
- Process parameters are established using decoupled molding techniques
- Once validated, historical data is recorded. This information identifies normal operating conditions, and can be used to identify changes within the molding process
- Process control limits are established to identify when change has occurred, and steps need to be taken for correcting the change to maintain process consistency.
These three steps are fundamental to processing strategies, and failure to complete the steps fully hampers a molder’s ability to quickly decipher changes that occur, identify root causes for the change and implement corrections that return to the original molding condition. With these steps clarified, we are now able to consider potential changes and how to use historical data as a tool.
We must first define what process characteristics should be recorded as key measurables. Here it is important to note that recording too much information is impossible. The more processing data we save when our process is in good running condition, the easier it becomes to identify changes and their root cause. The following measurables would be considered the minimum of measurements to be tracked:
- Fill Time
- Peak Pressure
- Screw Rotate Time
- Cycle Time
- Part Weight
At a minimum, these variables are the most common information recorded for comparison when molding conditions change. In most cases, they offer key data that can be used to identify change. Here is a list of other key measurables that can increase the amount of data available, and further define potential changes to the processing schematic when variance falls out of control:
- Water pressure (GPM/LPM, supply and return)
- Melt Temperature (measured in running state)
- Mold Temperature (water temperature, setpoint vs. actual)
- Mold Temperature (steel temperature, measured at various points while the mold is in a running state)
- Barrel temperature (Actual vs. setpoint, steel temperature recorded between bands with press in running state)
- Back Pressure (setpoint vs. actual)
- Hot runner temperature (this data is recorded with hot runner exposed, measuring steel temperature between bands.
- Moisture Analysis (based on manufacturer’s recommendations)
- Material Let-Down (measured upon receipt of each new material lot)
- Dryer (setpoint vs. actual, manual indicator on both supply and return hoses)
This additional data increases our ability to identify root cause for changes that occur within the molding process. Here are some examples of using combined data to clearly identify molding variances:
- During production, it is noted that a burn has started to occur. Upon reviewing data GPM of the return line has dropped, and mold temperature (steel temp) has increased. The root cause is identified as a plugged water line where the increased mold temperature occurred.
- A part is continuously sticking on a lifter. Upon review of thermolator data, it is discovered that the actual temp is 20F lower than recorded data. The root cause is identified as a valve stuck open
- Short/ unfilled parts become a scrap issue. Upon reviewing melt temperature and barrel temperature actual data, it is noted that the melt temperature is low, and actual barrel temperature is lower than previously recorded in the metering section. Maintenance is notified to inspect bands in the metering zone.
- A nylon part develops an issue with splay. Barrel and mold temperatures were reviewed, and all appears normal. Through moisture analysis, it is discovered that the material moisture is at .01, despite the material manufacturers recommendation of .10. The material is deemed overdried, and dryer throughput/ temperature/ operation are evaluated.
- A new lot of material is received. Upon review of the material let-down data, there is a significant difference between normal data and this specific lot. After further investigation, it is discovered that the material manufacturer mislabeled and sent the wrong material. The mistake was identified long before the material was used and misdiagnosed at the press.
- A process starts exhibiting problems with splay. Barrel, mold and thermolator temperatures are all consistent with historical data. Material moisture appears higher than normal. Upon review of historical data for the dryer, a significant drop in temperature on the dryer return line is identified using historical data. Maintenance is called to investigate the return side performance of the dryer.
The examples given above are just a few ways that increased historical data can help to improve troubleshooting capabilities. Increased data recording vastly improves a molder’s ability to pinpoint changes within the molding process. Faster, more complete evaluations of process variations will greatly improve the rate of return to a stable, normal process.