But don’t obsess over the distribution.

Yes, I said it. Normal data are nearly never normal.

In Six Sigma classes we study outliers, shift, drift and special cause events. But what we don’t always consider is that these “unexpected” data points may be part of the process and not as rare as we think.

First, let’s look at a set of screw torque data. The chart in FIGURE 1 is for a set of screw torques taken sequentially from a “smart” driver. We can see the data are normal (p=0.895), and the histogram and time series plot back that up.

There are more solder balls to inspect, so use oblique views.

With board area often at a premium in today’s sophisticated and high-performance electronics, using the third dimension of height to enable greater device functionality is ever more attractive. In particular, package-on-package (PoP) devices have been around for years and are popular for this purpose. From an x-ray inspection point-of-view, the most common POPs can be considered as two BGAs stacked on top of each other. Therefore, the real and “virtual” issues associated with BGAs, as mentioned in previous columns, also come into play for these devices.

However, we now have an additional dimension (sic) to consider in the features seen in the x-ray images, as well as just having a single BGA and anything within, or on the second side of, the board. On top of these, we also have the potential for overlap of solder balls from the two BGA layers (FIGURES 1 and 2). As with BGAs, all the interconnections are underneath the package, and are not visible optically, permitting x-ray the opportunity to make nondestructive inspection analysis. With twice the number of BGAs in the same place, does this mean twice the number of potential problems or more?

Could vendor collaboration get to the root cause of an intermittent soldering problem?

Simply put,  the ultimate function of a PCB assembly line is to create millions of solder joints without error. This task is complicated by the myriad materials that come together during assembly, and relies on the quality of each lead, pad and sphere to be soldered. When a soldering defect is discovered, it is common practice to presume the soldering materials are the cause, which seems logical, considering it is a solder defect. This assumption is often misplaced. This scenario plays out regularly, as illustrated in a recent case submitted to our failure analysis team for diagnosis.

In this case, the assembler had an intermittent solderability issue with a component. It brought the problem to its local representative’s attention on several occasions. The issue was isolated to a single component and was repairable at the rework station. It was a nuisance but didn’t interfere with production schedules. The situation was difficult to address; it was intermittent and subtle, but persistent. And each time the representative brought its solder supplier’s field engineer for site visits, the solderability issue was not present. Solder companies detest these types of issues because they have a negative effect on customer satisfaction and product perception.

A poor stencil aperture seal may lead to solder bridging.

When we think of a gasket, we often relate it to something experienced in our home, perhaps a plumbing gasket or a gas pipe. A tight seal – or gasket – ensures no leaks. In SMT printing, the same theory applies: A good gasket means no solder paste leaks and a better printing result. With stencil printing, gasketing occurs when the aperture cut into the stencil is pressed firmly against the corresponding pad so it forms a seal. To achieve a good gasket, the aperture should be at least 1:1 with the pad size. If the aperture is larger than the pad, a poor gasket will result. Our company generally advises a one mil (25µm) reduction of aperture size to ensure a tight seal. So, for example, if the pad is 125µm x 125µm, an aperture of 100µm x 100µm is recommended. All things being equal, this approach will help form a good, tight gasket.