Estimated reading time: 8 minutes

How to Achieve the Apex of Reliability

The biggest conference in the electronics industry will be upon us before you know it, and when you live in the Midwest, a trip to San Diego can’t come soon enough. So, in advance of IPC APEX EXPO 2019. I wanted to put together a quick snapshot of a few test methods in IPC TM-650 that are the most important to use regarding quality and reliability.

Before I get into specific test methods and their importance, let’s discuss IPC and TM-650 test methods in general. IPC was originally formed in 1957 as the Institute of Printed Circuits. A lot has changed since then, including what IPC’s acronym stands for; now, IPC is known as Association Connecting Electronics Industries. While IPC’s new name may not directly relate to the three letters they are known for, it more accurately describes the mission and industry in general. This was a good move because the scope of the IPC has drastically changed since 1957, as has the electronics industry.

The IPC has done an excellent job keeping pace with new advancements in the industry with test methods and educational offerings. In my position at a testing lab, we test samples of all kinds to different approved IPC test methods that are all found within TM-650, which is a big book full of testing that covers pretty much every conceivable aspect of manufacturing electronics. TM-650 had its first release in April of 1973 with around 200 test methods; today, it includes 253 active test methods and 18 old test methods. Current test methods are on a five-year review cycle within the 7-11 task group that updates them as needed for various reasons. The old test methods are still applicable but are not part of a review cycle.

Some of the current test derivatives of military specifications and others are original test methods designed by professionals from different corners of our industry based on specific materials or processes. The test methods are the backbone of the J-STD series, which are the assembly documents of choice that cover bare boards, assemblies, flux, paste, solder, etc. If you have any questions about how to assemble a product, you can find a section in one of those standards; within that, you will more than likely find a reference to a test method.

There are even standards that apply specifically to space and automotive industries among others with demands outside of most electronics. These standards are used in most assembly houses around the world to ensure quality and reliability. On any print, you should find a reference to what specifications to build to and the order of precedent as agreed upon between user and supplier (AABUS). IPC specs are also typically called out. That is a normal default that in lieu of direct specifications from the customer, an assembler should use IPC standards as assembly guidance documentation that will include tests from TM-650.

The test methods I want to discuss in this column are mostly related to cleanliness and different ways to determine if the process is clean enough for the intended end-use environment. Knowing the effect of residual ionic content is among the most important data points when looking at reliability. It is so critical because the ionic content is directly related to electrical leakage and electrochemical migration-related issues in a normal field service environment.

When you have the fire triangle—contamination, available atmospheric moisture, and bias differential—the risk of electrical leakage is greatly increased. There are no industry limits regarding allowable amounts of ionic content determined by ion chromatography because there isn’t a single set of cleanliness limits that apply to all electronics. It is important for each assembler to determine if the ionic content present will be detrimental to the field performance of your specific product.

The single test used by the industry the longest to test cleanliness is TM-650 2.3.25 Resistivity of Solvent Extract. The problem with using this test for acceptance is that it was designed in the 1970s and was never intended to be used for this purpose. You can find a very detailed white paper—IPC WP-019—that describes the history of this test method and why it should no longer be used as it most commonly is today. In short, the product being built when this test was introduced did not use bottom-terminated surface-mounted parts using no-clean solder paste among dozens of other material differences.

When addressing cleanliness, the most accurate method to determine the ionic content is ion chromatography. The test method for that is TM-650 2.3.28 Ionic Analysis of Circuit Boards, Ion Chromatography Method. This method requires subjecting the sample to a mixture of isopropyl alcohol and deionized water at 80°C for one hour. This will bring surface contaminants that can become soluble into a solution that will be processed through the ion chromatography (IC) equipment. After an IC test is complete, you will have the exact type and amounts of each anion, cation, and a general weak organic acid total. Each of the materials that go into the assembly process will have a chemical signature that can be matched back to the IC results.

This information is crucial in determining the risk of electrical leakage. If you find elevated levels of ionics, you can look at the IC results and compare them to flux activators, plating chemistries, wash signatures, bare board fab, handling, etc. When you know the process or material that is causing elevated ionic residues, you can then address that specific process and optimize it to a point where it leaves minimal active residues. The standard extraction method is usually done on a full board where you must calculate the total surface area of the sample to include a population factor of 10% if components are present. This means the results are an average of contamination if they were evenly spread across the entire sample—front and back.

While this information is good, it leaves a few questions unanswered. There is a growing understanding in the industry that to better determine the risk of electrical leakage or other contamination related issues, you need to look at much smaller areas of the sample. This allows you to look at specific processes like wave solder, hand solder, localized cleaning effectiveness, etc. This type of extraction can be done in several ways—both manual and automated. When the IC is processed on the effluent, the number is much more meaningful than an average cleanliness number. When elevated ionics are found after a localized extraction process, you can then go back and compare to data from the PC fabrication, component, or flux residues. From my perspective, IC analysis is generally the best tool to use for determining the possible impact processing residues will have on reliability.

The next test method I want to highlight is TM-650 2.6.3.7 Surface Insulation Resistance (SIR). This test is used for material and process qualification utilizing unpopulated test coupons with known spacing comb patterns in an elevated heat (40°C) and humidity (90%) chamber with constant voltage application. The test method was originally written to use the B-24 test boards, but in the past few years, more and more companies are using IPC B-52 test boards. This is a much better option because it incorporates components commonly used in today’s manufacturing.

There is another SIR test option—TM-650 2.6.3.3—which is similar but uses test parameters of 85°C and 85% RH, which can alter the chemical composition of weak organic acids and give false results. The test runs for no less than 72 hours with most running it to 168 hours. Measurements are taken every 20 minutes to monitor any effect the elevated heat and humidity have when combined with the residues present on the sample. Much like the IC test, the idea is to determine if the residues present will facilitate electrical leakage and/or electrochemical migration. The acceptance criteria details for this test are found in the J-STD-004 standard—section 3.4.1.4.1.

While SIR is a good test to look at the assembly house’s ability to process the chosen set of materials with their equipment, it does not necessarily translate to the actual product being built. For that test, I recommend temperature, humidity, and bias (THB) testing. This test is designed to accelerate electrical leakage or dendrite growth, particularly on device die surfaces that would never be tested with standard SIR testing. One standard set of parameters is 85°C and 85% RH for up to 1,000 hours, but that can vary.

THB testing is done on actual assemblies and not test coupons, which means there are no comb patterns to judge the results and you will use active components instead of dieless. There isn’t a related IPC test method for this, but it is commonly done in many sectors of the industry. You will need to create a test fixture to apply normal operating power and cycles with a monitor of expected feedback for a measurement point. While there isn’t an associated IPC test method, it is certainly a good idea for first article assemblies.

The main tests I would recommend for determining baseline cleanliness levels and what effect they will have on end-use reliability are IC, SIR, and THB tests. However, there are a handful of others that are worth the time and money. Cross-sectioning per IPC-TM-650 2.1.1 is a great way to get a better understanding of the solder joint quality as well as the PC fabrication process. Knowing there is a good intermetallic compound (IMC) will go a long way in determining if you will have possible issues with intermittent connectivity. A strong solder joint will also help with vibration issues in more severe operating environments.

If you are planning to use conformal coatings, it is a good idea to test how resistant to moisture it is using TM-650 2.6.3.4 Moisture and Insulation Resistance. This test uses 20 cycles of various temperature and humidity levels to help drive as much moisture as possible into the coating and see how effective it is at preventing ingress onto the surface of the assembly.

I’ve mentioned six test methods, so that leaves just 247 more active tests you could use to determine the reliability of your products. Did I mention I work for a test lab? Call me; let’s talk.