Predicting failures is vital for packaging technology


Some time ago, I wrote an article on what I call the “10 commandments of packaging” – my list of key things to keep in mind when it comes to developing new packaging solutions. packaging. Today, packaging engineers face more variables than ever before, from new substrate materials to more types of packaging with a greater range of complexity. With this complexity comes new challenges, and I felt that some of these commandments needed to be deepened, starting with “Anticipate Failures”.

Predicting failures requires knowing what to look for and having a methodology in place to help you anticipate failures so you can prevent them from happening. If you wake up one morning and see water on your kitchen floor, it’s clear you have a leaking hose. You need to fix the problem as soon as you see the first signs, or it will become catastrophic.

Having a mechanism in place to truly predict packaging process failures can make a huge difference in allowing us to stop them before they happen. A key example is the optimization of the resistance of the wire link. Wired cabling is an automated process, but the links must be tested to ensure tensile strength. Bonding material may build up on the tool, or the tool may experience wear and residue built up during the bonding process. Visually inspecting wire links is not a practical way to determine problems that may lead to link failure.

One way to alleviate this problem is to perform extraction tests and use statistical process control (SPC) techniques to build predictive models. By recording the tensile strength of the wire link on each product (i.e. the amount of force that can be applied before a link fails), we can create a database of thousands of pulls that the SPC software can analyze in order to locate anomalies. The use of SPC is pretty standard. Some companies just chart their process and display the print on the factory wall so they can see how they are doing.

The best use is to examine and analyze the data to understand what it is telling you over time. For example, a graph can show that the average pulling force was 14g in June, 13.7g in July, 13g in August, and 12.9g in September. This downward trend in pulling force tells us that something is wrong. SPC plots the dots and creates an analysis of that trend. You can then go back to the wire connection process, and you might find, for example, that you need to change the capillary more often.

Solving short-term problems is called a containment action – basically, a sort of band-aid. Unless you get rid of the root cause, the problem will persist. Techniques such as SPC – also known as Statistical Quality Control (SQC) – are essential to achieve this. W. Edwards Deming knew this. Considered by many to be the “father of quality,” Deming was a statistician and business consultant whose emphasis on statistical methods, continuous improvement and enterprise-wide quality formed the basis of this. which we now call Total Quality Management, or TQM. He went to Japan after WWII as an advisor to their census process, and his methods helped speed Japan’s recovery from the war and beyond. It has helped Japan, and ultimately the United States and the rest of the world, understand how to create processes that can be replicated in a uniform and consistent manner, so that you don’t waste time creating rejects.

Today, artificial intelligence (AI) and machine learning are integrated with SPC to accelerate learning and enable consistency in increasingly complex processes. The notion of lean manufacturing is an extension of SPC; being able to understand and assess variations and why they occur helps us to be more consistent and deliver what our customers need faster.

Reliability testing techniques
Before mass production, electrical packaging must pass reliability tests. The Thermal Cycling Test (TCT) is one of the standard reliability tests that has been commonly used in the electrical packaging industry. Ensuring that new products pass the TCT is a critical issue in the electronic packaging industry. Finite Element Method (FEM) simulation-based design technology can be used as a feasible development methodology for reliability assessment and reliability prediction of electronic packages.

Another method of detecting faults and ensuring reliability is to simply run the device for an extended period of time to detect problems. The High Temperature Reverse Bias Burn-In Test (HTRB) is a simple, inexpensive, and accelerated life test that helps sort faulty parts from the population. The main cause of failed units can be dielectric failure, conductor failure, metallization failure, etc. These failures are latent and manifest themselves randomly as device failures during the device life cycle. We use HTRB break-in tests to stress the device, speeding up these dormant faults to manifest as failures so that they can be filtered out during the infant mortality phase, thus preventing failures from occurring.

When something goes wrong with our business, customers ask for corrective action. We are able to generate a report that examines the “5M’s and E’s”: man, material, method, machine, measurement and environment. Anything that goes wrong is caused by one of these factors. Knowing which one created the problem allows us to determine how we should solve the problem. By evaluating these six factors at the start of a project, before if a failure occurs, we can step into a historical perspective that lets us know where we need to act to prevent a problem from reoccurring. This focus on continuous improvement and process optimization is vital for the future of advanced packaging.

Sam sadri

(All posts)

Sam Sadri is a senior process engineer at Quik-Pak.

Source link


About Author

Comments are closed.