Rectangular BGA pads are a bad idea.
Using rectangular apertures for round BGA pads or designing four corner pads can lead to a disaster.
Richard Stadem of General Dynamics said that printing paste outside of the pad is always a bad idea. After being printed on the pad and the solder mask, the fine solder balls have a difficult task of agglomerating into a single solder joint during reflow soldering. This is even more difficult if the pad is not defined in the solder mask, because then the beads must additionally bridge the gap between the solder mask and the pad edge.
As a result, solder balls that are not part of any solder joint can become trapped under the BGA. Moreover, they can go undetected, even during an X-ray examination, because they are often too small or outside the established X-Ray inspection zone.
Typically, X-Ray software is quite similar to AOI and some X-ray inspection machines are actually a combination of both (the terms AOX, AXI, or OAX are often used). The machine first locates the reference point (i.e. fiducial) and then proceeds to a programmed check routine, which is usually the first step of making a series of check measurements based on an offset from the base over a defined distance. In the second step, most commonly, each ball is measured based on its programmed pitch - its size and shape - and the voids are analyzed from the grayscale image.
This means that NO inspection is carried out in the areas between each target location. If individual particles or, worse, partially agglomerated solder balls are present in these 'orphan areas', they will not be detected by the automatic X-Ray. On the other hand, they can be large enough to exceed the minimum electrical spacing required between vias and BGA solder joints or between vias and adjacent conductive tracks or formed by solder connections and tracks.
As a result, the entire system may suffer from crosstalk as well as permanent or intermittent short circuits. Some of these free runaway balls can be removed in the washing process, but practice shows that they can never be removed. Moreover, if no-clean or RMA flux is used, literally hundreds of beads can remain trapped in its residues depositing beneath the component, causing real havoc.
Rectangular / square pads are also a problem.
There are claims in some internet users that may be interpreted that 'square pads and apertures can provide better paste printing results for BGA'. Richard Stadem strongly denies this: in this scenario, round BGA, CSP or flip-chip solder balls are placed on square pads with square paste deposits. Instead of a ready-made solder joint that should resemble a round vase on a round table, we get a connector that is round in the top half and connected to a circular pad on the bottom of the BGA, and has four corners where it connects to the PCB.
During the cooling down of reflow soldering, as well as during the entire life of the joint under temperature variation, the variation in the thermal expansion coefficient between the BGA and the PCB causes severe stresses to appear on the solder joints between them.
When tension is applied to the square pad under the BGA and if the forces are strong enough, the corners of the pad may be pulled upwards. This upward force breaks the bond between the pad and the substrate, often breaking not only the pad itself but also the paths connected to it.
This phenomenon is called 'cratering' (pad catering), and at the higher temperatures required for lead-free soldering and the more brittle nature of the solder joints themselves, this phenomenon is increasingly observed. This is a significant problem even with standard round pads, and the addition of four stress-concentrating corners has also been shown to increase the problem by an order of magnitude. We wrote about pad crateringu recently in a separate article.
Adapted from Richard Stadem of General Dynamics.
Cover photo: © InterPhone Service