How Can Molders Achieve Effective Process Monitoring?

Effective process monitoring is achieved through proper process development. To monitor effectively, you must ensure that the process is optimal and repeatable. Without repeatability, effective process monitoring will never be achieved.

Effective process monitoring does not simply mean setting upper and lower limits just before defects are created. This is what I call “defect limits.” Although it is important to know the “defect limits” of any process, these limits are not typically tight enough to be considered effective for process control.

Effective process-monitoring limits should be set so that if any abnormality is discovered outside of the normal process variation of the machine, it is segregated from good product. The normal variation of your process should be well within the threshold of the defect limits. The goal of establishing process limits should be to monitor the process that has been proven to produce acceptable product and not to create process windows. Wide process limits that try to create process windows within which process adjustments can be made can often cause more harm than good.

The thousands of interactions taking place in a molding process make it very difficult to prove out every scenario. For this reason, it makes more sense to establish a robust process and use process monitoring to ensure that this process repeats itself each and every time and that abnormalities are caught.

 

Determining Normal Variation

 

The process of identifying what is normal variation may differ depending on the product line, but I consider three 8-hr runs with complete shutdowns between each run to be a great start. It is extremely important to ensure no process changes are made during these runs, and you may want to undertake at least one with an alternate resin lot number. Process outputs should be recorded every 15 minutes, and the parts should be collected for that recorded process. Using this data, the effective process limits can be calculated, as long as all the parts collected are within specification.

You could set up stages of process monitoring so that any shot that is segregated would be considered suspect, if it is within the “defect limits,” until further review of that part can take place. This may not be necessary if your monitoring is set up in such a way that the machine will alert when scrap thresholds are reached. If so, the amount of product at risk will be minimal, and focus can then be placed on troubleshooting the issue at hand.

Process monitoring can be accomplished using the machine’s monitoring software; but in my opinion, third-party hardware is required for effective process monitoring. This will allow for independent verification of the machine outputs, as well as in-mold monitoring in the form of pressure transducers and thermocouples.

The ability to monitor the plastic and ensure that it reacts the same in the cavity from shot to shot is the ultimate form of effective process monitoring. We use the same methods for establishing the effective upper and lower limits for peak cavity pressure, pack rate, and cooling rate, as stated above. This establishes machine-independent, effective process monitoring, and regardless of the machine outputs, it can ensure the plastic was within established limits.

To monitor a process effectively, you must also have an effective means of part discrimination when out-of-process conditions arise. Automation options such as robots, reversing conveyors, and diverters can be used to accomplish this. If none of these is an option, the machine must be set up to shut down when any out-of-process condition is identified, and ejection must be prohibited, leaving the suspect parts in the mold. This is undesirable for several reasons, one of which being that a process interruption can cause instabilities in the process. Automating is the best option and will allow you to continue to mold, keeping the process stable. It is also critical to ensure you have built in effective reject logic in your monitoring to prevent the machine from running excessive scrap rates.

 

Failsafe to Avoid Failures

 

Is your process monitoring failsafe? Most monitoring systems are programmed to look for a reject signal. So, if the monitoring system doesn’t receive a reject signal, parts are considered to be within process limits. So, what happens with a system crash? Remember, these systems are basically computers, and we all know that even the best systems have software issues from time to time. If the system is only looking for reject signals, that system cannot be considered failsafe, and if the system locks up, it can continue to run until someone discovers the issue. Most of the third-party process-monitoring systems I have worked with will also accept a good-part signal if you make that part of your system specification. A process-monitoring system that requires a good-part signal to be received and defaults to reject, is a system that is truly failsafe.

It is important to remember that regardless of how tight monitoring limits are set, the first step in effective process monitoring is the process development. Establishing limits for unstable processing conditions or tooling conditions will ultimately prove ineffective in preventing unacceptable product from escaping. Once you have mastered process development, effective process monitoring ensures you’re always running to that process.