Reading time ( words)
A few years ago at Integrated Test Corporation, we found that the reaction plan for void fallout at electrical test was ineffective and not standardized. Like many PCB manufacturing facilities (including a Sanmina shop that I used to work at), the reaction plan consisted of cross section analysis to determine the void type.
Then, based on the type of void, we would either thermal stress and cross section coupons from passing boards or process passing boards through reflow simulation and retest electrical continuity for disposition.
In addition to these reaction plans, another that I have experienced at Sanmina included one where circuit boards would be processed numerous times through a micro-etch process and retested for electrical continuity if voiding was found within unfilled holes. If the panels withstood that, they should be okay, right? Unfortunately, none of these disposition methods are robust enough to ensure that vias with marginal connection are caught before shipment. As we all know, a few holes within a coupon is hardly representative of the thousands of holes within a circuit board, and processing production orders through a reflow simulation or micro-etch before assembly will negatively impact the life of that PCB.
Therefore, the only way to ensure that marginal products are not being shipped to the customer is to perform an electrical test on those suspect vias at a low enough resistance where minor differences between vias can be observed. A method that was evaluated and proven successful would be four-wire Kelvin testing; if characterized properly for your process, it can distinguish differences in copper thickness between holes. A failure discovered by this testing method is depicted in Figure 1.
Figure 1: Example of low copper discovered
At the time, Integrated Test Corporation did have Kelvin probes and a flying probe tester that was capable of performing the testing process. However, it had not been properly set up to accurately predict what the resistance measurements should be based on the aspect ratio of the via and the amount of copper in the hole.
The first few times the process was used, all that was accomplished was to indicate which holes had resistance measurements outside of the normal distribution of results. These were then sectioned, and it was found that they would have met the minimum copper criteria.
After a few instances of performing destructive analysis on nondiscrepant products, it was decided that this process required testing for proper characterization. A quick search of articles and white papers
yielded comprehensive descriptions of the process itself but not guides on how to set it up in production. Most of the papers available described building a baseline of resistance measurements with known good panels. Ultimately, the process that we desired was to accurately predict the resistance measurements based on drill aspect ratio and copper thickness. This would require a correlation between these measurements and a set of equations. These equations could then be used to set the maximum resistance specification during testing or to determine the copper thickness within plated through-holes without destructive analysis.
At Integrated Test Corporation, very high aspect ratio vias are common in production. So, we decided to design a test panel that was 0.300” that could be drilled with coupons including 0.010”, 0.012”, 0.015”, and 0.020” vias to characterize the process for aspect ratios up to 30:1. For each diameter, via coupons were included that were copper plated with 0.0002”, 0.0004”, 0.0006”, 0.0008”, and 0.001”, respectively. To ensure that the proper amount of copper plating was deposited in the holes, boards were processed through many different plating cycles, covering and uncovering coupons with resist during each cycle.
This was found to be more cost-effective than building panels exclusively for each copper plating thickness. However, doing the characterization in this manner would simplify the plating and imaging operations.
To read the full version of this article which originally appeared in the June 2019 issue of PCB007 Magazine, click here.